Foucault,Sellars,and the “conditions of possibility” of science

Abstract

Foucault and Sellars are representatives of conflicting philosophical traditions: whereas Foucault famously insisted that “power is everywhere,” Sellars proposed the well-known scientia mensura dictum. The tension between the two perspectives seems to be so strong that each of them ends up reducing the other to an epiphenomenal illusion. In this article, I shall attempt to show that the works of Sellars and Foucault are not necessarily irreconcilable. The common ground for this dialogue is what I shall define as a historico-practical conception of science. I will build this concept by tracing a connection between the Foucauldian notion of “conditions of possibility” of science and Sellars’s thesis about the “indispensability” of the manifest image. Finally, I will argue that this conception of science problematizes the clash between the scientific and manifest images of the world, paving the way for a different relationship between naturalism and critical theory.

The past three decades have witnessed a growing number of projects aimed at tracing connections between the analytic and continental traditions (Brandom 2002; Friedman 2000; Leiter 2002; Norris 1997; Prado 2003; Redding 2007; Rorty 1991; Rouse 1987). In more recent years, the idea of a pluralist approach that goes beyond the divide between these two philosophical traditions has made its way into the academic debate. Rather than trying to build bridges between two strands still conceived as essentially distinct, this pluralist approach foresees a future for philosophical thought and practice “no longer determined as ‘analytic’ or as ‘continental,’ but, instead, as a pluralistic synthesis of much of what is best in the diverse legacy of twentieth-century approaches” (Bell, Cutrofello, and Livingston 2016, 2). ¹ Despite these remarkable attempts to overcome the analytic-continental divide, dissention, and mutual distrust continue to characterize the relationship between critical theory (broadly constructed) and scientific naturalism.

Critical theorists have often accused scientific naturalism of forgetting the degree to which science is anchored in the framework of prescientific praxis. According to critical thinkers, this gesture of concealment does not only have epistemological consequences but also political ones, because it covers the sociopolitical structure on which science is rooted. ² In turn, philosophers embracing naturalism have argued that critical theory, and especially its post-structuralist declension, is invariably anti-scientific. ³ From this perspective, critical theory turns the scientific worldview into one of the many metanarratives in the history of Western culture: science becomes an epiphenomenal illusion of modern techniques of power, and the only relevance that critical theory attributes to it is that of capturing the ideological structure of modern societies.

Michel Foucault and Wilfrid Sellars are two illustrious representatives of these conflicting philosophical traditions. While Foucault famously insisted that “[p]ower is everywhere” (Foucault HS, 93), Sellars proposed the well-known neo-Protagorean dictum, according to which “science is the measure of all things” (Sellars EPM, § 42). The tension between the two perspectives seems to be so strong that each of them ends up reducing the other to an epiphenomenal illusion (De Caro 2018). It is not only the possibility of a pluralist philosophy that is lost but also the chance of a bridge between the two philosophical traditions. In this forward-thinking article, I shall attempt to elaborate an interdisciplinary and pluralist approach, to show that the works of Sellars and Foucault do not necessarily express two irreconcilable philosophies, but can help us to open a space of dialogue beyond the divide between scientific naturalism and critical theory. The common ground for this dialogue is what I shall define as a historico-practical conception of science. I will build this concept by drawing a connection between the Foucauldian notion of “conditions of possibility” of science and Sellars’s thesis about the “indispensability” of the manifest image. Finally, I will try to show that embracing this conception of science means to challenge the clash between the scientific and manifest images of the world, paving the way for a new relationship between naturalism and critique.

1. Science and its “conditions of possibility”

In this section, I will focus on the notion of “‘conditions of possibility’ of a science” (‘conditions de possibilité’ d'une science), as elaborated by Foucault in his 1968 essay “On the Archaeology of the Sciences: Response to the Epistemology Circle.” ⁴ In taking into consideration this text, I will also rely on the interpretation that Arnold I. Davidson provides of it in his 2001 book on historical epistemology and the formation of concepts. Davidson focuses on this essay to clarify the relationship between epistemology, archaeology, and genealogy; his choice does not appear accidental when we consider that Foucault wrote this essay during a turning point in his work. In fact, The Order of Things—the book on the archaeology of the human sciences—was published two years earlier, and the essay “Nietzsche, Genealogy, History”—in which he explicitly introduces the genealogical method—was to be released in 1971.

In the essay in question, Foucault argues that scientific knowledge, which in this and other contributions takes the name of connaissance, depends on specific conditions of possibility, where under “the general term of the ‘conditions of possibility’ of science, two heterogeneous systems must be distinguished” (EW II, 326). The first system is constituted by conditions “internal to scientific discourse in general, and cannot be defined other than through it” (EW II, 326). This “system” (système) defines the conditions of science qua science: “it is relative to its domain of objects, to the type of language it uses, to the concepts it has at its disposal, or it is seeking to establish; it defines the formal and semantic rules required for a statement to belong to the science” (EW II, 326). The internal conditions can be considered as the epistemological foundations of science, namely the norm-governed framework that defines the parameters and limits of scientific knowledge as such.

Moving to the second “system” of conditions, Foucault tells us that it “is concerned with the possibility of a science in its historical existence. It is external to the former system, and the two cannot be superposed” (Foucault EW II, 326). The second system of conditions is external to scientific knowledge and “is concerned with patterns that have their own consistency, laws of formation and autonomous disposition” (EW II, 326). According to Foucault, these patterns, or “discursive formations” (formations discursives), define a form of knowledge that he calls savoir. ⁵ Savoir is not a form of scientific knowledge, but this does not mean that it is “a rhapsody of false knowledges […] which the sciences, in their sovereignty, definitively thrust aside into the night of a prehistory.” If savoir is not the pre-history of a science, it should not even be thought of as “the outline of future sciences that are still confusedly wrapped around their futures” (EW II, 326). In other words, “knowledge [savoir] is not science in the successive displacement of its internal structures; it is the field of its actual history (EW II, 326). The external conditions of sciences constitute the norm-governed framework of savoir, and this structure is different from that of connaissance (scientific knowledge) and its internal conditions of possibility.

The main thesis that Davidson draws from Foucault’s essay is the following: “science stands to epistemology as knowledge (savior) […] stands to archaeology” (Davidson 2001, 193). In light of this conceptual structure, I would like to suggest that the external conditions of science are its archaeological conditions, while those which are internal form the epistemological conditions. If scientific statements can be given within the horizon defined by the first system of conditions, it is true that, at the same time, the entire domain of these statements takes place within a given form of savoir, which provides the external conditions of possibility for the existence of science and its internal conditions. Savoir defines a broader horizon than any given science; and shifts in the discursive ensembles of savoir can lead to transformations at the level of the structure of connaissance. Nevertheless, “in claiming that the discursive formations of knowledge (savoir) provide the (external) conditions of possibility for the (internal) structures of scientificity, one must be careful not to misinterpret the notion of conditions of possibility” (202–203). Savoir is not sufficient for the appearance of any form of connaissance, because it is a condition of possibility and not of actuality. Hence, scientific knowledge cannot be considered in any way as an epiphenomenon of its conditions of possibility.

Foucault is equally careful in clarifying that savoir does not have a uniform and unitary character. Discursive formations are patchy and relatively disconnected domains of knowledge. The notion of épistémè allows Foucault to account for the relationships that exist among different forms of savoir. In his The Order of Things, he resorts to the term épistémè, in a specialized sense, to mean the historical knowledge that connects the discursive formations within a particular epoch. ⁶ Thus, the épistémè is a “global configuration” that constitutes “the total set of relations that unite, at a given period, the discursive practices that give rise to epistemological figures, sciences, and possibly formalized systems” (Foucault AK, 191). A given épistémè does not define a unitary image, but an historically contingent “grid” or a “totality of relations” that resist any attempt to be fixed within a uniform kind form of knowledge or “type of rationality which, crossing the boundaries of the most varied sciences, manifests the sovereign unity of a subject, a spirit, or a period” (AK, 191)

This overview gives me a means of tracing, in the wake of Davidson's work, a link between archaeology and genealogy. Davidson notes that the chronological and conceptual distinction between these two methods should not be exaggerated. If genealogy was explicitly deployed in Discipline and Punish, it is equally important to recognize that “both Histoire de la folie and Naissance de la clinque were not only archaeological, but genealogical avant la lettre” (Davidson 2001, 204). In light of the problematization of the chronological relationship between archaeology and genealogical, I would like to challenge their clear-cut conceptual distinction. Indeed, in his 1970 inaugural lecture at the Collège de France, Foucault explained that archaeology and genealogy do not exclude one another, but they “are to alternate, support and complete each other.” (Foucault AK, 234). The introduction of the genealogical method should not be considered as a transformation aimed at affirming the obsolescence of archaeology, but as an enrichment of Foucault’s toolbox. Genealogy provides Foucault with the conceptual tools for working on aspects of historical investigation that would have otherwise remained included only implicitly in the archaeological perspective. The genealogical method allows Foucault to investigate the reciprocal connections between discursive practices and power relations. Hence, connaissance and its internal conditions are flanked by the power (pouvoir)/knowledge (savoir) relations. Returning to the question of archaeology, we can attempt to integrate Davidson's analysis by suggesting the following relationship: knowledge (savoir) stands to archaeology as power (pouvoir) stands to genealogy.

According to Foucault, power has a productive character: it constrains as it produces the social structure itself. It would, therefore, be erroneous to believe that power is given only in limited circumstances and with a repressive character. In fact, power pervades the whole social architecture and contributes to its formation in an essential sense. For this reason, Foucault, in a famous (or infamous) passage, tells us that “[p]ower is everywhere; not because it embraces everything, but because it comes from everywhere” (Foucault HS, 93). Since scientific knowledge takes place in society, it is not exempt from the conditions imposed by power, but is rather “subject to constant economic and political incitement […]; it is the object, under diverse forms, of immense diffusion and consumption […]; it is produced and transmitted under the control, dominant if not exclusive, of a few great political and economic apparatuses (university, army, writing, and media); finally, it is the issue of a whole political debate and social confrontation” (Foucault EW III, 131).

To better frame the relationship between knowledge and power, I would like to resort to the notion of dispositif. As Davidson reminds us: “Foucault developed the notion of a dispositif, an apparatus […], in order to be able to study the linkages, or network, that exists among elements within ‘a resolutely heterogeneous ensemble’,” namely “an ensemble consisting of both discursive and nondiscursive elements” (Davidson 2001, 205). When Foucault observes that a dispositif is a more complex entity than a discursive formation, he means that a dispositif is the system of relations that can be established through an “heterogeneous ensemble consisting of discourses, institutions, architectural forms, regulatory decisions, laws, administrative measures, scientific statements, philosophical, moral and philanthropic propositions–in short, the said as much as the unsaid” (Foucault PN, 194). Thus, the notion of dispositif complements that of discursive formation, including the latter, in a wider system of relations and power mechanisms. If, in his early writings, he calls épistémè the “global configuration” of discursive formations; in his later writings, Foucault will coin the notions of “ontology of actuality” (ontologie de l’actualité) to speak about the grid of knowledge/power relations that define a given historical context. ⁷

2. The historico-practical conception of science

Foucault’s archaeological and genealogical inquires have been interpreted by some of his adherents (and detractors), as a radical critique of scientific naturalism. When interpreted in this way, I believe that the critical-historical approach shows the historical dimension of any form of knowledge, but is unbale to account for the origin of history itself. In other words, “the omnipresence of power” (Foucault HS, 93) acquires a whole different meaning than the one I have previously provided, and becomes an omnipotent and ubiquitous ground which ultimately reveals itself to be another type of Absolute. This idea of the Absolute corresponds neither to a transcendent and all-embracing Mind outside and above finite human minds, nor to a transcendental Subject outside and above the empirical self, but it nevertheless shares the same scientific ineffability that characterizes onto-theological and onto-anthropological philosophies: power becomes a background condition of intelligibility that remains unexplained by empirical inquiry.

The historico-critical approach certainly risks taking this path, but, I also believe that another reading is possible, which questions an unconditioned scientism rather than science itself. Foucault is, in fact, careful not to erase the difference between connaissance and savoir. Therefore, it seems more appropriate to believe that his analysis is not aimed at disqualifying scientific knowledge as such, but to show that the latter does not give rise to a unified and perspectiveless image of the structure of the world. There are at least two compelling reasons why Foucault allows us to resist this conception of science. The first one concerns “science as science” (Foucault EW II, 326), namely connaissance as such. According to Foucault, connaissance is fragmented into a plurality of domains, which do not provide a unified theoretical image of the world. In his careful analysis of Foucault’s archaeology of science, Gary Gutting emphasizes Foucault’s “regional” (Foucault AK) conception of science, which eschews “grandly global theories” and shows the specificity of “particular disciplinary and chronological domains” (Gutting 1989, 53). Hence, the idea of a unified scientific image of the world does not do justice to the de facto plural forms of connaissance and their heterogeneous posits. According to Foucault, there is no overarching view that arrogates to itself the task of level out the various regions of connaissance. Hence, the conception of science that Foucault offers us seems to be that of a plurality of family resemblances, which generates some commonalities between the posits of the various scientific domains.

However, this conception of science still does not fully account for the originality of the Foucauldian perspective, and must be integrated with a second, fundamental aspect. According to Foucault, “knowledge is always the historical and circumstantial result of conditions outside the domain of knowledge. In reality, knowledge is an event that falls under the category of activity” (Foucault EW III, 13). If knowledge as such is an activity, then the same applies to scientific knowledge, which should be understood in relation to specific discursive and material practices. An understanding of science as an “activity” incorporates connaissance, and it also includes all the “conditions of possibility” that are indispensable for the production of connaissance itself. From this perspective, the idea that connaissance is possible independently of the norm-governed framework of scientific practice has lost sight of the fact that science is an activity conducted by a community of subjects. Hence, according to Foucault, connaissance depends on a given norm-governed framework-in-use, at the core of which lie the internal conditions. The attempt to eliminate or displace this norm-governed “system” would pull the rug out from under connaissance itself. The investigation into the internal conditions of science (such as the proper role of values in inference and assessing evidence) only scratches the surface of the conditions that make science possible. It is precisely by investigating the external conditions that contribute to establish the direction and applications of scientific research that Foucault can claim that science “isn’t outside power or lacking in power” (Foucault EW III, 131). As mentioned in the previous section, Foucault does not believe that these conditions mean the inevitable politicization of science. It is no coincidence that, already in his archaeological writings, he traces a distinction between the internal and external conditions of the scientific process, which implies that the internal conditions of scientific rationality has a degree of independence from the savoir and its conditions. ⁸

Although the existence of a conception of science as activity might seem obvious, incorporating this idea into the debate about naturalism has a profound impact, since it shows that the conditions of possibility for science are not the negative remnant of what should otherwise be an unconditioned scientific naturalism, but rather the indispensable prerequisite of connaissance itself. In other words, connaissance does not occur despite the norm-governed framework defined by the conditions previously examined, but because of them. I propose to call this conception of science the historico-practical framework of science. The criticism that, through Foucault, it is possible to move to scientism has an undeniable relevance for contemporary debates on naturalism and, in particular, for those visions of the scientific image as absolutely unconditioned. Among them could be included that of the American philosopher Wilfred Sellars, who has argued that “science is the measure of all things” (Sellars EPM, § 42). While, Sellars coined this neo-Protagorean motto, Foucault defended the omnipresence of power. Despite the apparent incompatibility between these two perspectives, in the next two sections, I will attempt to show that the works of these philosophers do not necessarily express irreconcilable perspectives, but can help us to define a conditoned-cum-unconditioned conception of science.

3. Sellars and the two images of the human being

In his 1963 essay “Philosophy and the Scientific Image of Man,” Sellars addressed what he considered to be the major problem confronting contemporary philosophy: the “clash” between “the ‘manifest’ image of man-in-the-world” and “the scientific image.” The two images delineate two ways in which modern human beings conceive the world, and their place in it: the image of the human being as a self-conscious agent bound by conceptual norms and that of the human being as a “complex physical system.” In his attempt to “join” the two images, Sellars is particularly careful in dispelling the belief that this “clash” is a conflict between a naïve and antiquated conception of human beings in the world, and one which is sophisticated and advanced. Instead, he defines it as the tension between “two pictures of essentially the same order of complexity” (Sellars PSIM, 4). On the one hand, exists the disciplined refinement of common sense through a perennial tradition of philosophical inquiry that began with “the great speculative systems of ancient and medieval philosophy,” and developed into the “major schools of contemporary Continental thought” and the “trends of contemporary British and American philosophy, which emphasize the analysis of ‘common sense’” (PSIM, 7–8). On the other hand, there is the methodical explanation of ordinary phenomena in terms of the imperceptible entities postulated by modern science, which developed from the scientific revolution of the 17th century, through subsequent chemical and biological revolutions, to the birth of contemporary physics and neuroscience.

According to Sellars, the primary aim of contemporary philosophy is to overcome this “crucial duality” through a comprehensive understanding of how the two conceptual frameworks can be joined in a single systematic account of the human-being-in-nature. To this end, he resorts to the analogy of “the stereoscopic vision, where two differing perspectives on a landscape are fused into one coherent experience” (PSIM, 4). In order to settle the competing claims of the two images, and reach a stereoscopic vision, Sellars distinguishes “the normative privileges of the former and the ontological rights of the latter” (Brassier 2011, 8). The ontological “primacy” of the scientific image is encapsulated by the well-known neo-Protagorean dictum, “in the dimension of describing and explaining the world, science is the measure of all things of what is that it is, and of what is not that it is not” (Sellars EPM, § 42). This motto expresses the possibility of a purely naturalistic explanatory account of the world. In principle, nothing is descriptively and explanatorily inaccessible to science, including the citadel of the mind. At the same time, Sellars does not believe that describing and explaining are the only “dimensions” of language use. Evaluating, prescribing, and negotiating, are indispensable dimensions of linguistic activity, in which science has no privilege (see deVries 2021).

Although the scientific image has ontological primacy when describing and explaining the world, the manifest one is nevertheless indispensable, because it furnishes us with the structure through which we can make sense of ourselves as concept-governed beings engaged in giving and asking reasons. And it is this susceptibility to norms that make us “persons” (or, to resort to a Foucauldian term, “subjects”) (see Brassier 2011). Thus, the manifest image provides the “categories pertaining to man as a person who finds himself confronted by standards (ethical, logical, etc.)” (Sellars PSIM, 38). Sellars deems the standards to which persons are confronted by to be essentially “social” (PSIM, 16), in the sense that normative notions exists only because they are taken as correct or incorrect by a community of persons. Hence, “[t]he fact that persons are beings bound up in a network of rights and duties, ultimately coincides with the fact that persons are members of a community because they can only be part of a network of rights and duties if there is a community of persons that takes those rights and duties to be rights and duties” (Levine 2019, 253). In other words, to think of “a featherless biped or dolphin or Martian as a person [or, in Foucauldian terms, as a subject] is to think of oneself, and it, as belonging to a community” (Sellars PSIM, 39), because the norm-governed space of the community “is the minimum unit in terms of which conceptual activity can be understood” (Sellars LTC, 512).

The manifest framework—through which human beings understand themselves as creatures engaged in discourses and practices involving normative notions, standards, rules, and intentions—is an indispensable dimension of human existence that, from Sellars’s perspective, cannot be reduced to the scientific conceptual framework; rather, it should be “joined” (Sellars PSIM, 40) to the scientific image in order to delineate stereoscopic vision. ⁹ Since the manifest image provides the conceptually irreducible and indispensable norm-governed framework that makes the scientific endeavour possible, the scientific image “does not stand on its own feet,” but is “methodologically dependent on the world of sophisticated common sense” (PSIM, 20). Although the manifest image is a necessary prerequisite for the development of the concepts and methods essential to the sciences, making the scientific image grounded methodologically on the manifest image, Sellars believes that there is no valid reason to grant the manifest image ontological priority over the scientific image, and so the scientific image is not grounded substantively on the manifest image.

The attempt to overcome the clash between the two images is extremely complex and ambitious. It may be that overcoming this dichotomy requires a partial revision of the very terms in which Sellars poses the problem. For instance, the scientia mensura dictum does not seem to belong to science but rather to philosophy: it is Sellars’s conception of philosophy that expresses and attempts to justify this judgment. Hence, the scientific image appears to be more connected to the manifest one and its perennial philosophy than what Sellars claims to beIn fact, the scientific image does not seem to configure a “view from nowhere” (Nagel 1986) that should be later “joined” with the manifest image, but, as deVries tells us, Sellars’s scientific image as “a detailed theoretical description of the structures of the world” (deVries 2012, 13) presupposes another conception of the scientific image “as a conceptual framework-in-use, a norm-governed framework that enables certain kinds of activity, namely, theory construction, and without which that activity would not be possible” (13). In other words, a theoretical scientific image presupposes a practical scientific image that makes certain “activities” possible. This means that any conceptual framework within which a theoretical scientific image of the world is possible “must already be a conceptual framework containing persons and the language of individual and community intentions” (12). These considerations allow deVries to problematize “the adequacy of the simple distinction between what is methodologically primary and what is ontologically primary, that Sellars uses to regiment the relation of the manifest and the scientific images” (13). The methodological dimension has an “implicit ontology”, which is “the ontology implicit in the language we use to deliberate about, and act within, the world” (15). It is interesting to notice that the conception of the scientific image, defined here by deVries, recalls the historico-practical conception of science that I have delineated, starting from Foucault’s analysis of the epistemological, archaeological, and genealogical conditions of possibility of science. In the next section, I will address the consequences of a reading of Sellars that focuses on his neo-Protagorean motto, and overlooks the historico-practical dimension of science. In particular, I will attempt to show that the attempt to eliminate the norm-governed framework of the community leads to a conception of the natural world haunted by anthropomorphic notions. In doing so, I will connect Foucault’s conception of science with his critical theory.

4. The ghost in the eliminativist machine

According to a consolidated definition of naturalism, “reality has no place for ‘supernatural’ or other ‘spooky’ kinds of entity,” and “the scientific method should be used to investigate all areas of reality, including the ‘human spirit’” (Papineau 2021, 1). Obviously, this broad conception of naturalism leaves plenty of room for debate about many questions, “such as what to include in the category of the supernatural [and] what we are to understand by ‘the sciences’ or a ‘scientific’ account of the human” (Macarthur 2010, 125). These questions also inform the debate about the relationship between the two images in Sellars. For example, let us suppose that the natural sciences are truly the measure of what is and, moreover, that they are developing a language for describing and explaining the world which is capable of replacing the grammar of the manifest image. At this point, the mental acts that constitute our common sense would be eliminated, because in the new scientific language there are no mental acts. Such an interpretation of the scientia mensura dictum raises the question concerning the “place” of those phenomena—such as meanings, morals, numbers and so on—which are both practically indispensable in our everyday lives, and yet absent from the world as it is understood by the natural sciences (De Caro and Macarthur 2004, 2010). The “placement problem” (Price 2011) is one of the founding questions of current debates about naturalism, and there is no doubt that influential forms of contemporary reductionism and eliminativism – questioning the legitimacy of common-sense phenomena in a scientific worldview – have been overtly influenced by Sellars. For instance, Paul Churchland, a former student of Sellars, maintains that “our common-sense conception of psychological phenomena constitutes a radically false theory; a theory so fundamentally defective that both the principles and the ontology of that theory will eventually be displaced, rather than smoothly reduced, by completed neuroscience” (Churchland 1986, 120). On this eliminativist view, the claims made by the manifest image are no more real than those about esoteric phenomena and spooky entities, meaning that they are destined to be replaced by a strict naturalist conception of the world. Although the stereoscopic vision shows that these interpretations of Sellars’s Neo-Protagorean motto miss the fact that “the manifest image is not overwhelmed in the synthesis” (Sellars PSIM, 9), it seems equally legitimate to argue that some of Sellars’s followers have adopted this or other forms of strict naturalism, and some of his critics believed Sellars was committed to eliminativism.

By considering these questions as purely academic, we neglect the influence scientific worldviews have on contemporary western culture. According to Jürgen Habermas, scientific naturalism represents, together with religion, one of the “two countervailing trends that mark the intellectual tenor of our age” (Habermas 2008, 1). Naturalism symbolizes not only the philosophical framework of leading scientific programmes in neuro- and cognitive sciences, evolutionary biology, and behavioral genetics but it also penetrates in society, giving shape to new forms of self-understanding of ourselves (see Habermas 2008). In tracing a connection between these reflections by Habermas and Foucault’s work, Maurizio Meloni points out that contemporary scientific naturalism, in spite of its increasing importance as an academic field of research, “appears even more relevant today at the level of an ontology of ourselves” (Meloni 2011, 151). It is thus a matter of understanding the processes of “subjectivation” (Foucault) through which the contemporary scientific worldviews contribute to shape new forms of subjectivity and new social figures (Lemke 2013; Mills 2011; Rose 2007; Tarizzo 2017). This reflexive gesture—which responds to the social, cultural, and political challenges posed by the advances in contemporary sciences—can be considered as one of the characterizing dimensions of Foucault’s thought and allows me to connect his philosophy of science with his critical approach to the ontology of ourselves (Giladi 2020; Gutting 1989; Rouse 2015). In a late interview entitled “The Subject and Power” (EW II), Foucault explains that his inquiries were not seeking to comprehend the mechanisms of knowledge and power per se, but rather the ways in which subjects conceive of themselves through these mechanisms (or dispositifs). Looking at his archaeological and genealogical inquires, Foucault individuates their critical core neither in the question of knowledge, nor in that of power, but in the question of the subject (Rovatti 2008, 2009). Although every retrospective view has a partially artificial character, I think it is possible to read most of his works, from the early investigations on madness to the later ones on biopolitics, as an inquiry into the way in which the regimes of knowledge and power contribute to revise the terms through which we understand ourselves and the world around us.

If we adopt this critical perspective and look at contemporary forms of strict scientific naturalism, in particular, at its eliminativist declensions, I believe we can argue that the dispositif of eliminativism produces a radical “desubjectivation” (Foucault) or depersonalization of the human being. The eliminativist aim to conceptually clean up the normative dimension, through which the subject is constituted, produces a paradoxical form of subjectivation, insofar as it is achieved through a process of desubjectivation. When one embraces an eliminativist version of naturalism and considers the natural sciences (and ultimately physics) as the only legitimate disciplines worth taking seriously to define the place of the subject in the natural world, the norm-governed ontology of ourselves assumes the same status as a supernatural, “spooky” phenomena. To better delineate the de-subjectifying effects of the dispositif of eliminativism, I would like to borrow the words that O’Shea uses to describe Sellars’s eliminativism, and redeploy them to frame the question of the subject from a critical perspective. The dispositif of eliminativism fuels a de-subjectified subjectivation through which the subject experiences its norm-governed existential space as “strictly speaking false; that it is in principle if not yet fully in practice, to be replaced by the explanatorily superior successor ontologies of the emerging scientific image of the world” (O’Shea 2009, 193). From this perspective, it becomes possible to consider the sophisticated common-sense framework as “irreducible” but nevertheless false; “and this framework would be ‘reducible’ in the sense of being in principle replaceable by the ideal scientific picture of ourselves and the world” (193). By hampering the grammar of subjectivation, not only at the individual level but also at the collective one, the dispositif of eliminativism turns the norm-governed space of society into a spooky, illusory entity.

This diagnosis recalls recent Foucauldian studies on the character of contemporary dispositifs. For instance, the Italian philosopher Giorgio Agamben has pointed out that contemporary dispositifs “no longer act as much through the production of a subject, as through the processes of what can be called desubjectivation” (Agamben 2009, 20). ¹⁰ Obviously, desubjectivation still amounts to a particular form of subjectivation, but the constitution of the subject occurs in “spectral form [forma spettrale]” (30). The conception of the subject as a spooky entity is the épistémè that contributes to the formation of specific form of “biopower” (Foucault). The result is a dispositif centered around the ideas of desubjectivation and neutralization of the norm-governed framework of the manifest image. Within this framework, the subjective deficit, namely the spectrality of the subject, is not a happenstance or an undesirable by-product, but a central characteristic of the naturalization of the human being. In so doing, this process of naturalization eliminates from the scientific image socially, ethically, or otherwise normatively inflected dimensions of the manifest image.

In adopting a Foucauldian perspective to challenge eliminativism, I do not want to suggest that Foucault was advocating the sacrality of the subject. In a famous (or infamous) passage of The Order of Things, Foucault proclaims the “death of man” (Foucault OT, 373); however, his critique of the subject should not be read as a definitive repudiation. Foucault situates his critique of the Subject (with a capital “s”) in the wake of Nietzsche’s announcement of the death of God (2005, 420). Foucault is aware that there is no naïve overcoming of the onto-theology of the subject and its shadows. He rather invites the reader to conceive a different philosophical approach starting from a “Nietzschean experience” (l’expérience de Nietzsche) constituted “by means of a philological critique, by means of a certain form of biologism” (OT, 372). Adopting “a certain form of biologism,” Foucault clearly embraces the naturalistic task to storm the citadel of the human mind. In other words, he embraces Nietzsche’s quest “to translate human beings back into nature” (Nietzsche 2002, 123). However, naturalism cannot eliminate the norm-governed space of conceptual thinking and, with it, the historic-critical practice that Foucault calls, in this context, “critical philology.” From this perspective, the eliminativist death of the subject has gone a step too far, and its attempt to displace the norm-governed grid of material and discursive practices inevitably leaves a “spooky” remainder. This residue is the ghost in the eliminativist machine which keeps haunting a strict conception of scientific naturalism, since it provides the indispensable conditions of possibility for science itself. The outcome of this naturalization of the subject seems to lead to a paradox: the desubjectivation produced by eliminativism seems to go hand in hand with a process of anthropomorphizing of its view of nature. In light of these considerations, it would be interesting to trace a genealogy of the ways in which these spectral reminders enter, under the shape of “anthropomorphic vestiges” (Piasentier 2020), within an eliminativist account of the natural world.

5. Unconditioned-cum-conditioned naturalism

There is no doubt that, according to Sellars, “the idea that epistemic facts can be analyzed without remainder—even ‘in principle’—into non-epistemic facts […] is […] a radical mistake” (Sellars EPM, § 5). This consideration goes hand in hand with the condition that a strict naturalism—aimed at cleaning up the normativity of the manifest image without leaving any “remainder”—threatens the inescapable requirement for our capacity to identify ourselves as humans, which is to say, as persons. In fact, according to Sellars, “man is that being which conceives of itself in terms of the manifest image. To the extent that the manifest image does not survive [...] to that extent man himself would not survive” (Sellars PSIM, 18). At the same time, Sellars’s “omnivorous” (O’Shea 2007, 3) naturalism sets the premises for problematic forms of reductionism and eliminativism. If this is the drift to which Sellars's project seems to be exposed, Foucault’s thought appears to be affected by the opposite problem: his critique of science risks throwing out the baby with the bath water, by turning the conditions of possibility for science into an omnivorous framework, ultimately regulated by a conception of power that works as an unexplained explainer. Here, Sellars comes to our aid, saying that the manifest image cannot but resort to a scientific explanation to account for its origin and existence. Any attempt by the manifest image to explain its own genesis is “bound to fail, for the manifest image contains the resources for such an attempt only in the sense that it provides the foundation on which scientific theory can build an explanatory framework” (Sellars PSIM, 17). It is only with “the scientific image of man in the world that we begin to see the main outlines of the way in which man came to have an image of himself-in-the-world. For we begin to see this matter of evolutionary development as a group phenomenon, a process which is illustrated at a simpler level by the evolutionary development, which explains the correspondence between the dancing of a worker bee and the location, relative to the sun, of the flower from which he comes” (PSIM, 17). We can further elaborate Sellars’s point, and suggest that the impossibility of the manifest image to explain its own existence and origin, is one of the key reasons why the perennial philosophy is always at risk of collapsing into a form of onto-theological (transcendent), onto-anthropological (transcendental), or onto-historical (quasi-transcendental) idealism. This is also what happens when we interpret the conception of power of Foucault as an all-pervasive principle of intelligibility, whose origin remains outside the reach of empirical inquiry. In turn, scientific knowledge becomes a mere instrument of social control in the hands of the regimes of discourse and power.

However, this interpretation does not seem to grasp the complexity of Foucauldian thought. The “Nietzschean experience,” which establishes a completely different relationship between critical theory and naturalism, should not be considered as an extemporaneous reference in Foucault’s works, but as one of its constituent characteristics. In “Nietzsche, Genealogy, History”—the 1971 essay inaugurating the genealogical method—Foucault seems to go back to that “Nietzschean experience” when—taking his cue from a Nietzschean exhortation to think “[h]istorically and physiologically” (Twilight of the Idols, in Foucault EW II, 328)—he challenges what he considered to be two of the main philosophical “idiosyncrasies” of the Western tradition: the “complete denial of the body” and “the absence of historical sense” (EW II, 328). Readers of Nietzsche often face a dilemma. Was the hammer with which he deconstructed the onto-theological tradition, and its shadows, forged in the furnace of scientific naturalism or in that of historical criticism? In my opinion, Foucault considers the “Nietzschean experience” as a philosophical approach, according to which, scientific naturalism and historico-critical philosophy are not two mutually exclusive perspectives but the constitutive elements of a complex philosophical vision. Borrowing an expression coined by Nietzsche and which refers to a historical point of view capable of serving as a corrective to the otherwise myopic standpoint of scientific naturalism, I shall call the philosophical vision that Foucault draws from Nietzsche as “second sight” (zweites Gesicht) (Nietzsche 2011, 39). The Foucauldian “second sight” neither implies a paradigm shift from an obsolete historico-critical philosophy to an advanced scientific naturalism, nor suggests a philosophy of second nature that establishes a domain of human experience, whose special status makes it immune to scientific investigation. My hypothesis is that the “second sight” is a philosophical vision that blends an “instinct for history” (2011, 39) with naturalism. Precisely for this reason, it can be compared with what Sellars referred to as “stereoscopic vision, where two differing perspectives on a landscape are fused into one coherent experience” (Sellars PSIM, 4). ¹¹ In other words, I believe that the “experience” Sellars is referring to, the “stereoscopic vision,” shares some similarities with what Foucault called the “Nietzschean experience,” namely, “second sight.”

In concluding this paper, I would like to integrate the analysis on naturalism conducted so far, with an unexpected, yet important, clue concerning what Foucault may have meant by “a certain biologism.” In his personal tribute on the thought and life of Foucault, Paul Veyne writes that “Foucault declared that Heidegger had been important for him […] However, in my own humble opinion, he had read little more than Heidegger’s Vom Wesen der Wahrheit and the big book on Nietzsche, which was indeed important for him as its paradoxical effect was to make him a Nietzschean, not a Heideggerian” (Veyne 2011, 147). Heidegger’s book on Nietzsche clearly had an important influence of Foucault. In that monumental work, Heidegger addressed Nietzsche’s “biologism” at-length. Although a comprehensive analysis of Heidegger’s reading is beyond the scope of this paper, I would like to provide some insights which I believe to be crucial to an understanding of Foucault’s re-interpretation of naturalism. According to Heidegger, the Nietzschean “world image” (Weltbild) is an “unconditioned” (Unbedingter) scientific naturalism, “not only in general, and as a consequence of a harmless opinion he may have propagated, but according to the innermost will of his thought” (Heidegger 1991b, 40). ¹² Heidegger locates the conceptual starting point of Nietzschean naturalism in a passage from On the Genealogy of Morality, where the author wonders, “when will all these shadows of God no longer darken us? When will we have a completely de-deified nature?” (Nietzsche 2001; in Heidegger 1991a, 94). According to Heidegger, Nietzsche rejects all forms of transcendent metaphysics and their “shadows”—whether they are supernatural entities (such as God or the Platonic ideas) or supernatural properties of the human mind (such as the self-transparent pure intellect or the immortal soul). ¹³ Heidegger explains that Nietzsche rejects supernaturalism and its “shadows” through a commitment to a naturalism that “directly adopts and extends concepts and key propositions” from the sciences of his time (Heidegger 1991b, 45). Heidegger is careful to underline that Nietzsche draws distinctions between the scientific disciplines that inquire into “the realms of plant, animal, and human life” (1991b, 41). Therefore, the Nietzschean scientific image is characterized by an internal pluralism when it comes to the posits and methods of science. As already mentioned in section 2, according to Foucault, science is a multifaceted, discursive, and material practice, which is articulated in a significant number of fields of research, resorts to a great diversity of methods, and employs heterogeneous conceptual and ontological resources. Affirming such a “regional” conception of science means to question the very idea of a single, idealized scientific image, and with it, the notion of a stringent naturalism. Thus, Foucault seems to reinforce and broaden the scientific pluralism already outlined by Heidegger’s reading of Nietzsche.

Nevertheless, a “unconditioned” naturalism would contradict Foucault's conception of science's conditions of possibility, if it wasn’t that Heidegger integrates this thesis with another one. According to Heidegger, Nietzsche is aware that a commitment to naturalism entails “metaphysical decisions” (Heidegger 1991b, 45–6): when Nietzsche outlines the “unconditioned” character of his naturalism, he is not thinking strictly scientifically; “rather, he grounds this natural world-view metaphysically” (1991b, 46) or, we could say, philosophically. ¹⁴ The scientific naturalism that Heidegger extrapolates from Nietzsche’s works seems to lead to a contradiction: Heidegger claims that the Nietzschean scientific “world image” is “unconditioned,” and yet, at the same time, he insists that Nietzsche supplements the “unconditioned” nature of his naturalism by imposing on it some philosophical conditions. In other words, when Heidegger insists that Nietzsche grounds his scientific “worldview” philosophically, he is telling us that Nietzsche's scientific image does not stand on its own feet, but depends on that perennial tradition of philosophical inquiry, which, with Sellars, we can call manifest image.

Heidegger is careful to heed the plurality of voices that Nietzsche puts into dialogue with one another, in defining his scientific naturalism. His inquiry leads him to wonder whether Nietzsche was, ultimately, a scientific naturalist. According to Heidegger, the different perspectives interwoven into Nietzsche’s philosophy make it impossible to answer this question with a simple “yes” or “no” (see Heidegger 1991b, 47). To convey the complexity of Nietzschean naturalism, Heidegger juxtaposes two apparently incompatible positions with one another and concludes that Nietzsche’s work does not ultimately fall into the trap of biologism. I believe Foucault would concur with Heidegger on the impossibility of responding to the question of whether Nietzsche defends a form of scientific naturalism with a simple “yes” or “no.” This is why he resorts to the phrase “a certain form of biologism” to define his naturalism. Although Foucault’s decision to resort to the term “biologism” to define his naturalism may appear misleading—because of the negative connotation of this term—it nevertheless shows his debt towards Heidegger’s reading of Nietzsche. Foucault’s books dispel suspicions of a latent biologism, and they show that his naturalism is not only characterized by an internal pluralism but also acknowledges the indispensability of “critical philology.” From this perspective, Foucault’s philosophy could be defined as a form of critical naturalism. Despite the differences between Foucault’s and Sellars’s conceptions of science, if we ask whether Sellars defends a form of scientific naturalism, I think we could answer just as Heidegger did when he responded to the question about Nietzschean naturalism: whether one responds “yes” or “no,” one always gets stuck in the foreground of his thinking. ¹⁵

Obviously, much more would need to be said to make this conception of naturalism plausible, but I hope that I have done enough to show that Sellars’s and Foucault’s philosophies are not necessarily irreconcilable.