2.1 The Beginnings of a Dialogue. The Rhetorical Analysis of the Scientific Text

In this first chapter, we retrace some of the essential steps and junctures at the basis of an original research perspective, the aim of which was to lay the foundations of a meticulous investigation into the logics that determine the effectiveness of scientific discourse. In particular, as we will see, one of the fundamental assumptions of Latour’s proposal, relevant both in terms of theoretical reflection and methodological implications, is to claim the opportunity to adopt a multidisciplinary approach, combining semiotics and ethnomethodology.

The starting point of what will evolve over time into a full-fledged anthropological approach to scientific discourse is a text published by Latour and Paolo Fabbri in 1977 titled La rhétorique de la science: Pouvoir et devoir dans un article de science exacte which effectively constitutes the first article dedicated to the semiotics of scientific texts. The goal of the research project outlined in this seminal essay is to demonstrate that the textual forms of the exact sciences are not reduced, as commonly believed, to the cold and impersonal description of an experimental practice. On the contrary, they manifest the presence of a complex rhetoric at the service of a true pragmatic intent: to induce the recipients of the scientific community to recognize as true only the utterances expressed by the authors, while discrediting the adversaries active in the same disciplinary field. From this perspective, the scientific text is reconsidered as a device of power, a battleground involving various players, a space animated by conflictual moves that are anything but obvious and elementary. The commonly accepted perspective regarding the strength of the scientific text is thus overturned and is now attributed not to its (presumed) neutrality, but rather to the effectiveness of its construction.

In order to demonstrate how the reader’s perception of the objectivity of scientific discourse is based on a series of precise moves in terms of signification, the two authors aim to conduct a multidisciplinary analysis at the intersection of sociology of science and semiotics, selecting as a reference text an essay on neuroendocrinology considered “cutting-edgeFootnote 1 “and published in 1962 in the Comptes rendus de l’Académie des sciences of Paris.Footnote 2 Upon careful examination, the richness of the scientific article emerges from its surroundings, precisely in the threshold space of the text that Genette (1987) defines as paratext or more precisely peritext, which in the perspective of sociology of science is already considered a place crossed by a strategic action that extends over two levels: that relating to the author (choice of scientific discipline, title, timing of publication and obviously of the journal) and that relating to the institutions involved in various ways (authorization protocols, sources of funding, choice of the laboratory where to conduct the experiment).

The scientific text thus lends itself to being considered under a new light, that is as the object of a strategic positioning both manifest (precisely in the paratextual messages located inside it) and implicit, relative to the conflictual scenario that regulates the relationships between the subjects engaged in the same frontier scientific field. It is from this premise that the detailed study of the article begins, breaking it down and analysing it in its constituent elements and in all their concatenations, significantly employing the same conceptual apparatus used and consolidated in the study of literary texts.

Consistent with a structuralist approach to text analysis, the deconstruction of the article is indeed based on a guided procedure in which a first fundamental distinction is made between its pragmatic dimension (related to the authors, their motivations, the recipients to whom the text is addressed), and its textual one. The properly textual plane is in turn articulated into two levels: the enunciation and the utterance. The system of enunciation concerns the linguistic mechanisms through which the authors decide whether or not to clearly signal their presence in the article, opting for the use of the first person or the third person. The plane of the utterance is instead reconstructed taking into account the textual elements that do not explicitly refer to the subjects of scientific discourse.

A first relevant piece of evidence that the decomposition of the text allows to emerge concerns the fact that this example of a scientific article is anything but impersonal. In what appears for all intents and purposes as a report-narrative of a laboratory experiment two textual mechanisms play a decisive role: the modalizations and the anaphora. The first consist of all the utterances capable of modifying prepositions (“postulate”; “have already said”; “the conditions have not been collected”; “without prematurely concluding”) of which the article is widely disseminated. The anaphora consist in the numerous references of the text to other parts of itself or to other texts that preceded its elaboration. And it is precisely this last element that seems to play a decisive role in the construction of the argumentative effectiveness of the article. The examined text, indeed, it is not only studded with a significant number of references, but the presence of a very precise system of connections between some sections of the article is also significant, a real chain of links.

The numerous quotations in the text that define the perimeter of scientific discourse, that is, its context of reference, are anaphora. But anaphora do not only concern the verbal dimension of the article. What emerges is a complex structure of references where the quotations refer to different textual types (tables and diagrams) in which the data obtained during the laboratory experiment are reported using instruments that are responsible for attesting the validity of the researchers’ assertions, serving as documentary evidence. This apparatus of documents, which is commonly taken for granted, considered “natural” in the description of the procedures and results obtained by researchers engaged in frontier sciences, is defined by Latour and Fabbri as symbolic subtextFootnote 3 and plays a strategic role in the functioning of the article because it is on this that the demonstration is anchored, that the scientific fact takes shape.

Finally, another type of anaphora concerns the title and abstract which do not merely evoke the text but instead synthesize all the information.

The identification and analysis of the intricate structure of connections set up in the article thus allows us to clear the field of reflection from a misunderstanding. The interest that guides this novel approach to the study of scientific texts is not at all confined to the need to understand what makes the genre of scientific discourse peculiar in terms of language style. The main stake rather concerns a reflection on the logics that regulate the production and circulation of knowledge (and power). In particular, the discovery of a logic that regulates the multiplication of internal references to the text, without which this discursive genre would fall, forces us to rethink the classic problem of the referent, of the object of discourse, what it refers to.

In other literary practices, there is either a referent which is not textual or else there is no referent at all—as in fiction intended as such. In the article presented here, there is indeed a referent, but it consists of an accumulation of texts: the context upon which the article acts, the infratext upon which it is based, part B, upon which part A is based. It is as if the paper's solidity—some might say its objectivity—stems from the correspondences established by interleaving each of these different layers of texts. Beneath the scientific text one finds not nature, but the literature of instruments (Latour & Fabbri, 2000: 121).

As is evident, this is a decidedly relevant implication on the theoretical level. The scientific text reveals itself as a the textual object that is not at all linear, rather it appears as a heavily layered construct, scattered with argumentative devices, interwoven with internal and external references, to such an extent that instead of a presumed direct link with nature (understood as the final referent) an intricate network of other texts emerges: in other words, a complex rhetoric at the service of a persuasive type logic.

Moreover, the information transmitted by the researchers involved in the endocrinology experiment appears rather scant; a glaring proof of this is the massive use of technical terms, which, being widely shared by specialized readers, do not increase their knowledge in any way.

Consequently, the strength of this type of scientific article lies not so much in transmitting information, but rather in the ability to convince, in its agency, understood as action exercised towards a multitude of other subjects, internal and external to the world of laboratories: colleagues, institutions, sources of funding, public opinion.

The text thus reveals itself as a complex semiotic device, whose analysis overcomes a naive vision of the relationship between texts and referents, between science and nature.

Every text seeks to convince, but the scientific text achieves this effect by developing layers of texts consistent with one another which serve as mutual referents. Here, for example, bioassays, the stages of purification, the rules of the procedure followed, the statistical analysis—all are in agreement. This agreement is a break from the disagreements that pre- ceded it. Where before there were claims, now there exists an object (Latour & Fabbri, 2000: 122–123).

The presence of a nesting of texts that within the scientific article follows a precise order thus enables us to rethink the relationship between the scientific explanation and physiological phenomena in an innovative way. To speak of the rhetoric of science does not mean to invoke a metaphor to refer to the style of scientific discourse, to the surface of the text (its linguistic manifestation), but rather to rethink its overall functioning in terms of persuasive effectiveness.

The referent of scientific discourse is no longer given a priori but is understood as the outcome of a process, as the ability of the text to reify its own meaning, generating it progressively, to construct and gradually enhance the truth of the discourse. Here we clearly find one of the fundamental assumptions of generative semiotic theory, which consists in emphasizing the impossibility of resorting to an external referent, consequently postulating the autonomy and the immanent character of every language.

The investigation into the truth of the scientific utterance thus comes to be rethought, in the terms of the theoretical project of Algirdas Julien Greimas, as veridiction.Footnote 4

[…] due to the fact that it is no longer considered as the representation of a truth exterior to it, discourse is no longer satisfied with the simple inscription of the marks of veridiction. “Truth”, in order to be spoken and assumed, must move toward the domains of the enunciator and the enunciatee. The enunciator is no longer presumed to produce true discourses, but discourse producing a “truth” meaning effect. From this point of view, the production of truth corresponds to the exercise of a particular cognitive doing, of a causing-to-seem-true that can be called, without any pejorative nuance, persuasive doing (Greimas & Courtés, 1982: 368)

The transition from a traditional conception of reference to a dynamic, processual vision, focused on the analysis of the procedures through which the so-called referentializationFootnote 5 of discourse takes shape, plays a central role in the development of Latour’s thought which in the following years, as we will see, while recognizing the great importance of the Greimas’ perspective, will be characterized by the introduction of some shifts dense with implications (theoretical and methodological) for the observation of the signification at play in social phenomena.

An aspect that in any case emerges as central, from this first study, consists in noting how the rhetoric of the scientific text is functional to a polemical logic, a real “agonistics” in which the article acts exercising a persuasive force within the clash between a multitude of actors engaged through a series of disputes to assert the authoritativeness of their own work.

This conflict has specific rules, it unfolds according to codified moves that in the text take the form of precise linguistic traces: the modalizations. The utterance of the exact sciences, beyond stylistic issues, can take two different forms: the simple one and the modalized one. The first case, typical of successful scientific demonstrations, is that of affirmations that circulate in scientific discourse without any restriction (“A is B”). In the second case, instead, the assertion of a subject is modified by the intervention of an antagonist who through a new declaration questions its reliability. A typical move of polemical action thus consists in reversing an assertion as happens in the following passages: “One of us has expressed reservations about the conclusions of Shibuzawa”, “There were not present all the conditions necessary to confirm that the active fraction of Schreiber et al. acts only …”.

In light of these considerations, Latour and Fabbri propose, therefore, to radically rethink the functioning of scientific ideology, polemicizing with the perspective, common in philosophy, of considering the latter as a sort of theatrical staging in which the backstage is hidden to show the public a theoretical dissertation devoid of plot and characters. On the contrary, the operation at stake in scientific discourse never consists in hiding the mechanisms of its production but in highlighting them, in exhibiting them to the point that the representation can be rethought as the history of the conditions of its realization. In this sense

It is even possible to define the frontiers of a science as the place where opponents are constantly forcing assertions (énoncés) back into the experimental conditions under which they were produced. Any “cold” science, in contrast, is presented as a sequence of affirmations—at least until a new front is established, remobilising assertions (mobiliser de nouveau les énoncés) and revealing their true origin (Latour & Fabbri, 2000: 124).

The stakes of the rhetoric of science are therefore very large and strongly desired by the subjects involved in research. It is indeed about the authority (to be built and claimed in the text), from which descends the credit of the profession of scientist and ultimately the power to generate and circumscribe (temporarily by definition) a new field of study, precisely a frontier.

The gain obtained from the use of a novel multidisciplinary approach to the scientific text thus materializes in two different types of transformation.

The first is a movement of extension and consists in expanding to an article of exact science the conceptual framework developed for the analysis of literary texts.Footnote 6

The second transformation is defined as inversion and consists in reversing the perspective rooted in common sense regarding the relationship that exists between nature and scientific research. In fact, while scientific production is commonly considered a speculative activity directed towards nature, in this new perspective it is rethought as a set of actions primarily directed towards a field of research—that of frontier scientific literature (intertextuality). The effectiveness of scientific demonstration is thus revealed in the presence of a military-type logic in which “[…] nature provides the ammunition whereby offensive strikes are made invincible” (Latour & Fabbri, 2000: 130).

2.2 For a Morphology of Scientific Narrative

If from a chronological point of view the semiotics of the scientific text is inaugurated by the publication of this article, the ambitious research project aimed at understanding the logics that ensure the rhetorical effectiveness of the discourse of exact sciences takes off and finds inspiration in the work of Françoise Bastide, precisely in a thematic seminar held in 1977–1978 linked to the seminar of “General Semantics” directed by Algirdas Julien Greimas and focused on the analysis of a text by Claude Bernard.Footnote 7

Bastide, an original figure of a researcher capable of combining her training as a physiologist with the methodological rigor of textual analysis developed in the field of structuralist-oriented semiotics, has played a central role in understanding the so-called operative character of scientific texts, that is, the mechanisms of formation of their pragmatic effectiveness. The premise from which her work starts is that the discourse of experimental sciences, in the typical codified form of the scientific report of a laboratory experiment, is an “object” of exemplary analysis for the semiotics of the text because its meaning would reside in the presence of an articulated structure composed of a series of distinct levels placed in sequence.

Bastide’s proposal thus consists in rethinking the scientific text dedicated to the laboratory experiment no longer as a descriptive utterance but rather as a particular form of narrative text, that is, as a semiotic construct whose elements are arranged in relation to each other according to a path that goes from the simplest to the most complex, from the most abstract to the most concrete. In this research perspective, it is important to clarify that the adjective “narrative” is not used in a metaphorical sense but rather in a technical sense to designate the morphology of the scientific text, as Bastide herself specifies, motivating the choice to examine Bernard’s article.

[…] Bernard, when dealing with phenomena of the natural world, proceeds by constructing increasingly complex narratives in successive steps, gradually inserting actors, times and locations around the main program of sugar formation in the liver. Moreover, he generates at each step the different possible narratives to account for a state or an action. The discourse of experimental sciences therefore seems particularly favorable to the study of conversion procedures between levels of meaning relevance (my trans.; Bastide, 1979: 10).

One of the most relevant points in Bastide’s work consists in demonstrating how the persuasive capacity of the scientific article is based, as highlighted earlier, on the presence of a real stratification (feuilletage) of text levels, an ordered overlap of expressive planes concatenated among themselves according to a precise sequence. At the first level of stratification, easily identifiable, and which consists of title, subtitle, abstract and notes, a plurality of other components are added that manifest the presence of distinct semiotic systems that transcend the dimension of verbal language. These are the indispensable equipment of every scientific explanation: equations, tables, images, photographs. The stratification of these different planes can reach levels of intensity such that in the history of science, as Latour (2009) himself reminds us, it becomes the indicator of the scientific word, even to the point of being considered synonymous with science.

This plurality of heterogeneous elements is all the more significant the more it strategically allows the subject of the scientific discourse to disseminate within the text a series of traces that refer to the practice of the experiment in the laboratory, strengthening in the eyes of the recipient the effect of reality. Under the magnifying glass of semiotics, the meticulous analysis of the text allows to bring out a further element that characterizes the functioning of the scientific discourse and that consists in the delegation by the figure of the scientist to a large number of non-human actors (the technical instruments) that play a decisive role in the argumentative effectiveness of the text, manifesting a series of skills and visualization strategies that are entirely peculiar. Since the essential function exercised by the equipment involved in the scientific experiment is that of making visible what is invisible to the human eye, it is possible therefore to rethink the scientific article as a real visualization device. What makes the non-human actors involved in the procedure of the experiment (and in its report) relevant for semiotics is the observation that, beyond common sense that tends to consider their functioning neutral, they contribute to exert towards the enunciatee (the scientific community) a persuasive action, acting as mediation elements in the chain of transformations that allow the transition from “nature” to the demonstration of the “scientific fact”. As Bastide specifies:

The enunciator does not hide at all, as in fantastic stories, but projects into the utterance-text a "paper" enunciator who stages himself as a witness of a phenomenon, of an event, and tells it to an enunciatee who was not present, to convince him of the "reality" of what he has observed. The process of persuasion goes through the transparency of the operations of the change of substance of the expression, when what is "seen" is transformed into discourse. It is for this reason that we find simultaneously choices made at the level of experimental devices that allow to "see" […] and a cancellation of the enunciator in front of the "facts" that he has collected, and that show by themselves what is worth seeing (my trans.; 1985a).

Following this perspective, the scientific result, that is the performance of science, thus presupposes a doing (the experiment), which in turn implies the possession of a competence (knowing-how-to-do) that tactically the human subject of the scientific practice delegates in part to a potentially wide series of heterogeneous non-human actors.

The process, in its minimal articulation, is depicted in the following way:

A schematic exhibits the alignment of letters with arrows indicating their sequence, E leads to O, O leads to S, S leads to Op, Op leads to P, and P leads to finally R.

At the extremes of the scheme, we find nature (E) and the recipient (R), that is, the researcher/scientific community. (O) indicates the invisible structure of the natural phenomenon under investigation, which through the intervention of an operator subject (S Op) is transformed into a product (P) whose structure is visible.

Following this perspective, the “scientific fact” comes to be rethought as the product (P) of a process of mediation that involves potentially very extensive and complex chains of agents (human and non-human). The idea of a direct correspondence, without mediation, between signs and things is thus definitively overcome, the scientific text is never transparent (but can certainly seem such, and this is mainly its persuasive vocation). Nature can only be grasped, therefore, through a sequence of operations of translation. The reference thus gives way to a referencing strategy.

Despite everything, traces of this process remain and the originality of Bastide’s work is further revealed in the choice to “interrogate” the functioning of the visualization device, focusing attention on the contribution that technical objects make in visualizing the “scientific fact”, in particular on what is commonly considered as the direct testimony of a real phenomenon, like a simple imprint without mediations: the scientific image. Bastide’s position once again aims to overturn common sense and consists in supporting the thesis that the visual apparatus widely used in the texts of the exact sciences can be considered in all respects as an iconography.

Clearly, photography is one of the types of images used most frequently in scientific texts and it is therefore primarily on this that the effectiveness of semiotic analysis must be measured. First of all, it is necessary to note that “a ‘scientific’ photograph, publishable in an article, is the complete opposite of a family father’s photo showing the child, the cat, the landscape, and maybe even the bicycle” (my trans.; Bastide, 1985b). The first peculiar characteristic of scientific photography consists in reducing as much as possible the information to be conveyed, in channeling the meaning of the photographed object in order not to generate confusion in its recipient. In this type of snapshot, the so-called polysemy, the coexistence of a multitude of meanings that characterizes the aesthetic forms of the photographic image, is thus lost. The sense of scientific photography lies rather primarily in its functionality, in making visible what is necessary to reinforce the argument of the text, taking the place of the human actor in the research process.

Referring to the scheme illustrated earlier, it can be observed how scientific photography allows for the reduction of intermediate stages between the object of nature “O” (intensional) and the object of the article “P” (extensional), particularly by eliminating as many stages as possible where the human actor explicitly reveals himself/herself. On a theoretical level, however, this does not in any way mean considering this type of images as belonging to the category of indexical signs, in the sense given by Peirce; that is, the meaning of the photograph does not at all resolve into a direct relationship with the referent, in the presumed ability to testify without mediation the presence of the phenomenon investigated in the experiment. Rather, what appears particularly relevant is the presence of a complex symbolic apparatus whose presence derives from the modes of production of the image itself. In the space of a laboratory, highlighting an “object” through the photographic tool necessarily implies acting on the plane of reality, particularly through two processes: selection and contrast.

The first procedure consists in framing the photographed object, in selecting only what is considered relevant, erasing from the image what Bastide calls the “background noise”.

The contrast, on the other hand, acts within the framing and is a procedure of highlighting the photographed object with the aim of detaching it from the background, of highlighting only the elements considered relevant to make visible what the human eye is not able to perceive.

The actions of selection and contrast are relevant because they contribute to orient the reader’s interpretation of the scientific image, activating mechanisms that semiotics defines as semi-symbolic and which consist in producing signification by associating differences on the plane of the signifier with differences on the plane of the signified. Thus, the classic mode that consists in opposing “light” vs “dark” (category of the signifier) can be used to bring out the object from the background, highlighting, for example, the composition of a muscle tissue. The dimensions at play in the constructionFootnote 8 of the photographic image are in any case numerous and range from the use of colour, to transparency, from the position of the elements in the framed space (high/low, central/peripheral to the cases that recall explicitly the coded model of the geographical map) to their orientation, up to the work of highlighting the state of the surface, its “texture” (e.g. smooth/granular).

The idea of scientific photography as a “natural” element is thus definitively put into crisis, not only thanks to the analysis of the internal mechanisms that regulate its functioning, but also by observing the relationship it has with the verbal dimension of the text. The photographic image, understood as an element of a unitary visualization device (the scientific article), in fact always participates in a double movement that goes from illustrations to text and vice versa, acting as a guarantee of assertions and also ensuring the linearity and homogeneity of the scientific document. The scientific image works like the article, with which it mutually reinforces (through referentialization) and with it participates in the same persuasive logic acting as “[…] a tactical maneuver, a no-way-out ambush in which, if someone tries a reading of the results different from that carried out by the authors, the change of direction is blocked by a specific argument” (my trans.; Bastide, 1985b).

This tactical manoeuvre is ensured in the text by the presence of a rich iconographic apparatus of which photographs, however relevant, are only one of the components. Traces of the role played by non-human actors in the visualization device also emerge in the use of graphs and tables, in whose functioning a series of conventions comes into play, the more relevant the more their presence is disguised by a rhetoric of persuasion aimed at celebrating the scientific illustrations as arguments that do not allow the possibility of reply.Footnote 9

It is important at this point to clarify two theoretical issues that arise in the pioneering study of Bastide and that, as will be seen later, will take on relevance in the work of Latour and more generally in the field of so-called Science and Technology Studies.

The materiality and the specific practices of use of the non-human agents involved in the construction of scientific facts and their narration inevitably impose a constraint on the possible representations of the results. For example, Bastide recalls, among optical devices, the electron microscope, thanks to a trick called “shading”, can be used by researchers to produce the effect of height, or to make it “as if” the object were detached from the background.

The presence of photographs, graphs or tables in a scientific text is therefore also a condition of the choice to use specific equipment during the experiment. In other words, what is relevant for the analysis of the rhetorical devices of persuasion at play in science is not only the dimension of discourse (the text), but also that of the material conditions that ensure its production, the tangible characteristics of technical objects and the practices of their use. A second issue, connected to this passage, concerns the dynamic, procedural conception of scientific “facts” and their discourse. If on the one hand scientific texts, like literary ones, can be considered in all respects as constructed texts, whose effectiveness takes the form of a complex rhetoric, on the other hand the former respond to a peculiar logic. The traces of scientific doing that stratify in the articles of experimental sciences, giving rise to even very long concatenations, always assume a tactical relevance in the sense that they are exhibited by the subjects of the discourse as evidence capable of guaranteeing the reliability of the results achieved. In other words, the traces act in the text as memory devices of the phases that make up an experiment and as such must be able to be retraced, a posteriori, by the scientific community to evaluate its solidity, its strength.

If photography is so widely used in the sciences (and is at the same time subject to meticulous controls and frequent disputes) it is also because it fulfils a precious function of archiving, guaranteeing a record to which access must be allowed at any time.Footnote 10 The traces are then evidence that allow us to connect the text to what is distant in space and time (the natural phenomenon intended as the object of research) marking a difference with the forms of literary narrative.

Venturing into the analysis of scientific discourse thus forces us to go beyond the limits of the text understood in a literal and traditional sense to investigate the network of relationships that involves multiple spaces, actors and times (laboratories and scientists, specialized journals, scientific dissemination, etc.). The criticism of a naive vision of reference therefore brings with it the rethinking of a central dichotomy in classical semiotic theory and in the social sciences: that between the text and the context.

2.3 Does the Scientific Image Not Exist?

In the years following the publication of Bastide’s pioneering research, the study of the signification at work in scientific texts continued to emerge clearly in the work of Latour,Footnote 11 who in 1987 published Science in action.

This is a work, destined to arouse considerable interest, which sets out to explore science in the stages of its very construction. In particular, Latour’s analysis focuses on the conflictual dynamics at work in the laboratories, the role of scientific literature in legitimizing discoveries and their creators, and the far from linear ways in which inventions progressively become accepted and assimilated, contributing to redefining even the shared idea of nature. By reconstructing the vicissitudes underlying the invention of the diesel engine or the mass diffusion of the Eastman Kodak instant camera, Latour shows the central role that controversies play in the emergence of scientific facts, their dissemination and their pragmatic repercussions, highlighting how the stabilization of disputes in science cannot be traced back, as common sense would suggest, to the evidence imposed by nature (the last word is up to the phenomena themselves) but to a network of multifaceted actants, human and non-human, material and immaterial, to a “long heterogeneous list of resources and allies that scientists were gathering to make dissent impossible” (Latour, 1985: 103).

In Latour’s view, therefore, science is action and the sequence of transformations on which its effectiveness is based must necessarily respond to a principle of reversibility. If a first movement consists in projecting the events of a laboratory experiment into a new space (the scientific text), whose narrative form responds to a tactical need (to convince) by resorting to rhetoric, on the other hand, the strength of the text, and its argumentative force require that it be able to withstand a second mirror movement, allowing the sequence of inscriptions (evidence) on which the demonstration is based to be retraced. Otherwise, the argument is a fallacy. In this sense, the perspective advanced by Latour is anything but relativist, rather it can be defined as a form of “realist constructivism”, insofar as he does not intend to deny the existence of reality, or argue that facts do not exist, but rather to convince us of the fact that “‘out-there-ness’ is a consequence of scientific work, rather than its cause” (Latour & Woolgar, 1986: 182).

The rethinking of the reference as a dynamic type of concatenation focused on transformation operations that link the text to the conditions of its production and circulation carries with it two significant consequences from a theoretical and methodological point of view. The first, as we have seen, consists in the abandonment of an anthropomorphic vision of agency, the second in overcoming the distance that separates the text from the conditions of its production and circulation. While regarding the first point, the affinity with semiotic epistemology appears solid because it is centred on the notion of the actant, the second would be more problematic. In this regard, we anticipate here some considerations that will be expanded upon in the following chapters.

As is well known, in the sociological sense, the notion of context refers to the scenario that circumscribes the space within which the actions carried out by human subjects unfold. Conversely, in the perspective advanced by Latour, if a being (human or non-human) is (intentionally or unintentionally) involved in the concatenation of transformations that make possible the emergence of a social phenomenon, it must be considered in all respects as an actor and not as part of the context. In this sense, there are no elements involved in a course of actions that cannot be considered as actors; the discriminating dimension rather concerns their ability to exert an influence in the unfolding of a phenomenon. The only elements that must be ignored in the reconstruction of the forces at play and their concatenations are therefore those that prove to be non-influential. In this sense, the context only concerns the elements that are marginal, external to the unfolding of a phenomenon, to the point that the context is reduced to everything that is irrelevant.

Now, returning to the comparison with semiotics, the critical issue would arise precisely from the analysis of the concatenations that in the emergence of a scientific phenomenon involve human and non-human actors. Traditionally, in fact, semiotics conceives the text as a narrative construct built from a basic enunciative operation, called débrayage (disengagement), which allows the subject of the communicative act to project “outside himself/herself” the three fundamental categories of any discursive activity: space, time, and subject. Consequently, every utterance, even the one that apparently seems more impersonal and “objective”, implies an enunciation and manifests its traces, internal elements to the text that must be distinguished from the concrete figures involved in the actual realization of the communicative process. In this sense, it is therefore necessary to underline the distance between the empirical subjects that are outside the text, that is the sender and the receiver, and their textual simulacra, defined as enunciator and enunciatee. The débrayage is defined as enunciational if it consists in the projection within the utterance of the simulacra of the subject of the enunciation (first-person discourses, dialogues); it is instead defined as enunciative if it consists in installing in the text subjects different from those of its enunciation (objectified discourse, in the third person).

Commonly, in the development of a text, the operation of débrayage is accompanied by an inverse movement, of return, defined as embrayage (engagement). As seen previously, these two discursive regimes often materialize in a text in the form of progressive encapsulations, giving rise to effects of reality, since each previous level constitutes a referential plane with respect to the next. However, one of the assumptions of Greimas’ semiotic theory is that the space separating the text from its author and its reader can never truly be bridged. Total embrayage in other words is impossible to conceive, as this would imply the (impossible) erasure of every trace of discourse.Footnote 12

It is precisely from this point that Latour’s criticism begins, specifically from the observation that unlike what happens in a literary text, the force of scientific texts (their rigor), presupposing the possibility of retracing the chain of traces (inscriptions) left by the transformations carried out during an experiment to verify its validity (evidence), requires that the actors of the narrative withstand the pressure of a form of return which is particularly binding and that cannot be ignored. Returning to the work of Bastide, and on the still open questions that it prefigures for a semiotics of scientific texts, Latour (2009) highlights, therefore, how the question of the internal referent would represent a potentially problematic turn for the study of signification, due to the centrality assumed by the literary text within the theoretical and methodological framework of the discipline.

This need for return is due to the fact that the finger of colleagues who read an article points to a picture within a text and demands in a certain way that the person responsible for this picture shows himself. This aspect establishes the extremely original character of this situation, which we generally do not find in many literary fields. The fundamental point, therefore, that makes the question of the usefulness of semiotic tools a bit more uncertain, relates in some way to access to what is distant. The need to produce information by accessing phenomena that are inaccessible, because they are too distant, too small or too old, requires that the characters that populate scientific texts be subjected to a pressure, a need for return, of re-embrayage that characterizes the chain of construction of scientific conviction and does not fall into categories easily framed in a semiotic model (my trans.; Latour, 2009: 255-256).

The comparison with the chain of translations on which science in action is based thus leads the scholar to affirm that the scientific image does not exist. With this polemical stance, Latour intends to reiterate the difficulty of analysing a scientific image in itself, that is, detached from the complex network of translations of which it is a necessary component. If, as we have seen, the study of the effectiveness of the scientific text allows to abandon the idea of a direct anchoring between the words of science and things (the “nature”) bringing out the complex work of translation and stratification of the visualization device set up to account for the object of research, what becomes important are the traces produced by the equipment used to measure its operation, the inscriptions with which they are projected from the space/time of the experiment in the laboratory into the space of the text (article).

The meaning of these inscriptions must then be sought beyond their similarity (iconicity) with the represented object or their physical proximity to the object (indexicality). Rather it consists in the ability to produce and exhibit evidence capable of strengthening the assertions of the text, of contributing to create a conviction (the “scientific fact”), through the representation of a series of transformations that regulate the transition from an unknown phenomenon to a known one, or from an unstable to a stable one.

From this point of view, the scientific image, understood as a broad category within which different types of illustration (photography, graph etc.) fall, only makes sense in function of the referential concatenation of which it is part, of the relationship that is established with the other components of the scientific process, whether they are human (the scientists) and non-human (technical instruments). Observing scientists at work and describing the disputes at the foundation of scientific discoveries thus leads Latour to assert that the scientific photograph, isolated from the argumentative structure of which it is part, is not able to exercise a higher deictic functionFootnote 13 compared to the other forms of inscription that compose the visualization device in its entirety.

Thus, the image is never an endpoint, but only one of the deictic elements within the text, whether we are talking about the image as such or whether it has as its only purpose of simplifying some perceptual judgments. With the scientific image we are dealing with an object that is completely outside the problem of images and that allows to introduce a difference between the characters and the resources of semiotics. We could therefore study the characters of the story, but also the phenomenon, quite different, of the involvement of these characters in the referential chain, where the latter—as we have now understood—does not mean "external reference", but continuation of the referential chain of the text within the laboratory or within the scientific community (my trans.; Latour, 2009: 254–255).

A single scientific image can therefore be considered significant only in function of the transformations it participates in during the argumentation process, otherwise it is deprived of its referential force due to the selection procedure with which it is obtained, essentially equivalent to a freeze frame.

Examining scientific phenomena as dynamic practices centered on traduction or translation mechanisms thus allows us to highlight a distinct attribute that sets them apart from traditional narratives. In scientific practice, there is a phenomenon as peculiar as it is relevant that Latour calls immutable mobiles, and that can be summarized as follows: something, which is not of the order of content but of expression, is preserved through transformations. This expression, deliberately contradictory, is introduced to try to account for a phenomenon as complex and problematic as maintaining something constant through a series of modifications of the plane of expression.

To describe this phenomenon, which we will explore in detail in the next chapter, we take the example of a technical object like the pedocomparator, a tool for comparing soil samples, used by scientists who study soil composition (Latour, 1999). This technical object consists of a box divided into compartments marked by an identification code into which the soil samples taken by scientists within the selected area are inserted. The function of the shelf is to trace a map of the geological composition of the investigated region, allowing scientists to subsequently perform a sequence of analyses (sending the samples to laboratories thousands of kilometers away) that finally translate into a set of images, tables, and graphs. This wooden frame containing a series of little cubes allows information about the analysed soil to circulate, making it possible to move the soil samples, but at the same time ensuring their preservation. The pedocomparator thus reveals itself as a very particular scientific object, a “concept-object” or a “fact-construct” capable of reconciling reality (the organic matter of the soil) with its codification, the plane of matter with that of the form used for the purpose of understanding its composition, the articulation.

The transition from reality to its scientific description is not therefore conceived as a leap but as a sequence of small intermediate and reversible steps capable of producing a series of transformations that, taken individually, appear to be of little relevance but in the set they form, they manage to bring about a change on the ontological plane, from matter to representation.

The functioning of the referential chain therefore leads to a rethinking of the role played by the notion of text, here considered only as one of the events that make up a transformation movement that consists in the circulation—translation of objects of scientific discourse. A significant role is rather assigned to the plane of experience, to the practices that make possible the construction and control of the investigated phenomenon.

The limit of a semiotics of scientific phenomena would then consist, according to Latour, in a theory of the sign based on the signifier/signified pair developed from the study of narrative texts.

This device is not useful for the study of scientific activity, in which the fundamental problem is not at all that of the signifier and the meaning, but that of a very strange relationship with the text, so far inexplicably foreign to the philosophy of sciences, namely the maintenance of a constant through the modifications of the plane of expression (my trans.; 2009: 261).

According to this perspective, Latour asserts, Bastide’s work would be even more significant because it would testify to the apparently paradoxical coexistence of two competencies. The semiotic one, put into practice by meticulously analysing the scientific text in all the levels that guarantee its functioning, and the scientific one, strengthened by the experience accumulated in the field, in the years spent in the laboratory.

Latour’s challenge and invitation to address Bastide’s pioneering work on the status of the scientific image prompt contemporary semiotic research to respond,Footnote 14 as evidenced, among others, by the work of Dondero and Fontanille (2014), aimed at exploring the semiotic dimension inherent in the functioning of scientific images linked to a variety of disciplines. In their research, the two semioticians openly recognize the relevance of Latour’s work on the scientific practices at play in laboratories, agreeing that “a photograph in itself, taken in isolation, cannot be used from a scientific standpoint” (Dondero & Fontanille, 2014: 131) and reiterating, at the same time, that the scientific character of the image “derives from the connections established, from a mediation assigning semiotic content purely to relations between objects” (ibid). However, one issue remains open, that of the invitation to overcome the distinction between text and context, to test the strength of the theoretical framework of semiotics in the space that unfolds outside the boundaries of the consolidated formats of textuality.

The next chapters are dedicated to the tension and constant confrontation between these two impulses.