Some epistemological problems of the idea of “scientific practice” as a historiographical tool


Dr. Godfrey Guillaumin


II Workshop on the Notion of “Scientific Practice”

Instituto de Investigaciones Filosóficas, UNAM

México, D.F. June 14-16, 2004


There are many interesting and important differences between the

various scientific disciplines. Paleontology or meteorology

may be as interestingly different from one another

and from high-energy physics as they are from macroeconomics

or political science. All have been shaped by a history of

internal development and interaction with other scientific

fields and other social practices.

(Rouse 2002, p. 95).


1.- Introduction

            Two important issues in the current philosophy of science are the topic of “scientific practice” and the theme of “normativity in historicism”. There are different philosophical projects that try to articulate a satisfactory notion of scientific practice, while others are looking for the best way to preserve normative claims derived from an empirical record such as the history of science. My recent work in the philosophy of science has been an attempt to articulate an acceptable proposal of the second task. Nevertheless, I think that a satisfactory notion of scientific practice could be helpful.

            Explaining how the change of epistemic concepts in science has normative effects among a community of researchers, and attempting to explain through historical concepts the justification of epistemic normativity are two interrelated issues not yet sufficiently clarified. Nevertheless, developing a historical account about how epistemic ideas were in fact changed in the past implies an important challenge, namely, how historical studies of epistemic ideas might articulate a normative dimension. In other words, do the facts of the past connect with what ought to be done in the future? When we take seriously the idea that historical studies of epistemic ideas require the preservation of a normative dimension, we discover two main problems: Firstly, how non-normative claims become normative through their history and, secondly, how their normative power spreads among the pertinent scientific community. The first question points to the genesis of the normative character of epistemic ideas in science, meanwhile the second attempts to explain their normative scope.

            Among many philosophical and historical studies of science, the so called, “naturalistic fallacy” still continues to represent a challenge. The problem is how can we describe or explain an historical episode within the history of science and at the same time preserve normative force? Larry Laudan has argued this idea and he has affirmed that: “One can show that a thoroughly ‘scientific’ and robustly ‘descriptive’ methodology will have normative consequences” (Laudan, 1996, p. 133). I agree completely with Laudan´s claim, but the problem is how exactly we can develop a historical analysis of science that is at the same time inherently normative? In other words, how should a historical epistemology preserve a normative dimension?  What analytical tools should we use in order to achieve this historical/normative enterprise?

            What I want to show here is that this project could employ a deflationary notion of practice in order to maintain its historical and normative character. Recently, Sergio Martínez (2003) and Joseph Rouse (2002) have developed interesting proposals that take very seriously the normative dimension of practices. Because I have insufficient space here to discuss the ideas of both authors, I’ll concentrate my analysis mainly on Rouse’s work. Joseph Rouse’s most recent book, How Scientific Practices Matter (2002), provides an interesting notion of practice, which could be useful for my approach to historical epistemology. Nevertheless, as I will show, I have some criticisms of Rouse’s conception of practice, particularly his understanding of nature and normativity. I start my presentation by showing in general terms how an historical epistemology must be normative. Then, I will explain why it might be useful to use a deflationary notion of scientific practice as an analytical notion, based on Rouse’s discussion about scientific practice as a normative notion.


2.- Normative character of historical epistemology

            There is not consensus among different projects on historical epistemology about how we exactly understand historical epistemology.[1]  Ian Hacking has affirmed that the expression “historical epistemology” was originally coined to express a concern with very general concepts that have to do with knowledge, belief, opinion, objectivity, reason and so forth. According to Hacking, recent works on historical epistemology, such as Daston (1991, 2000), Galison (1992), Porter (1995), etc., have indicated that epistemological concepts are not constants—free-standing ideas that are just there, timelessly (Hacking 2000, p.8). Nevertheless, “historical epistemology” is an awkward expression. On the one hand, at least for a philosophy of science inherited from logical positivism, epistemic ideas were analyzed by formal analysis because this philosophical approach to the study of science mainly looked for some sort of a priori or formal structures of thought. Among authors of this tradition, epistemic notions such as validity, objectivity, rational judgment, etc., could not be explained on the basis of empirical matters of fact. Additionally, the meaning and justification of empirical scientific claims were also thought to require a nonempirical grounding for their authority, which means that the normative character of science was not empirical. Logical or necessary structures were the means to secure the normativity of meaning and justification of the empirical content of science; hence, the normative character of epistemology has no historical dimension. Therefore, under this philosophical tradition, the expression “historical epistemology” could indicate just a descriptive- historical study of past epistemic ideas without a normative weight, and if someone seeks to find a normative character to historical epistemology, that would be a contradiction in termini. One the other hand, if we emphasizes the historical character of this expression and say, as Hacking does, that what historical epistemology analyzes is the historical character of epistemological concepts, then we are making just a historical-descriptive analysis of epistemological ideas. In this sense, we have projects such as many of Berlin’s that pay special attention to the integration of social, cultural, and cognitive studies of science. According to my point of view, an historical epistemology must preserve a normative dimension because if not, then it would be either just an epistemology that illustrated its epistemic statements via the history of science or simply a history of science without a normative dimension. The philosophical problem is, thus, how an historical study of epistemic ideas should be a normative enterprise. In the end, this difficulty belongs to a species of normative naturalism that I have elsewhere named “historical normativism”.

            In order to face this problem, it is necessary to answer two central questions, firstly, what is a naturalistic analysis of scientific knowledge? And secondly, what exactly does it mean for an historical study of epistemic notions to preserve normativity? One traditional answer is that instead of merely describing an historical episode, we compare it with another historical episode in order to establish which aspects of both episodes were better, or more progressive, etc. than others. For instance, instead of merely describing Newton’s physics; we can establish whether this episode was more progressive than Descartes’ or Galileo’s physics. Clearly, in order to carry out this strategy, we need criteria to determine scientific progress, the problem, then, is: where does this criteria come from? Are these criteria in some sense a priori? What we must historically analyze in its historical context is which theory was successful and why it was superior to its rival. We should not impose our evaluative criteria to judge the past, as in Lakatos’ rational reconstructions. The crucial and very important analytical point here is to not forget that typically scientific controversies take time to establish which theory is better, for instance, during many decades following the publication of Newton’s Principia there were controversies regarding in what sense Newton´s theory was better than Descartes. My point is that an historical epistemology, in order to be normative, must proceed in two steps. The first step is to analyze the historical reasons, proofs, rules, empirical evidence, etc. (all these epistemic notions) through which finally, after years of controversies, a scientific belief about nature or methods was considered better than another and, secondly, to establish whether the arguments of this episode were present in another historical controversy. The second step has the aim of looking for normative arguments beyond a particular historical case and tries to establish a sort of geography of epistemic normativity through different areas.  In short, normative criteria for science are taken from their own historical development, not from some sort of a priori arguments. In this sense, historical epistemology is a naturalistic enterprise.

            One central idea of this naturalistic and normative project of historical epistemology lies in the following simple issue: epistemic concepts, such as truth, proof, demonstration, probability, and so on, change through their history. Historians or philosophers describe and explain these changes, but, and this is the important point, these changes generate epistemic normativity for scientists within their field. A change of epistemic concepts cannot take place without a change of their epistemic normativity relevant to the field of inquiry. It is well known among many philosophers and historians of science that the distinction between “inside” and “outside” in science is difficult to trace, there are even extreme authors that argue that this distinction is totally undesirable, however in this case this distinction is completely crucial, because the normative consequences of a change of an epistemic idea directly affect scientists or researchers in a specific area, but not the philosopher or historian who studies or explains this change. What I want to stress here is that normative force of epistemic concepts affects the practice of scientists. It is common in the history of science that some authors of an area take epistemic ideas from another area, so when I use the idea of “inside” I do not refer to cases of transportation of epistemic ideas from one area to another. Normative consequences of conceptual changes locally shape some activities of inquiry in specific areas. In other words, changes of epistemic notions affect the epistemic normativity of the practice where these changes have taken place. Thus, historical epistemology must focus on analyzing not just how and why an epistemic idea changes but also, and more importantly, it must focus on elaborating a map of the normative contours which generated the epistemic ideas.

            The notion of “practice” is useful in this project because the normative consequences of epistemic controversies are clearly detected through the notion of practice, as we shall see further. For the moment, I want to concentrate on the analysis of the notion of “epistemic normativity” that can be formulated in different ways, but I understand it as follows: cognitive normativity in science is the way in which scientific inquiry usually associates certain cognitive values with specific epistemic ideas and the way in which scientific inquiry prescribes (or prohibits) specific actions in order to obtain specific cognitive results of the inquiry,(which  are usually discoveries). For instance, at the beginning of the seventeenth century, in early Natural Philosophy, “experimental demonstration” was associated to the idea of knowledge without doubt. However, decades later, Robert Boyle did not talk about experimental demonstration at all. For him, experimental knowledge provided just probable knowledge.

            There are two significant aspects of this formulation of cognitive normativity that are necessary to emphasize. On the one hand, it refers to beliefs as well as actions of inquiry, and on the other hand, such associations and prescriptions are historically generated. This formulation of cognitive normativity also has an important social aspect that can be formulated as follows: cognitive normativity is expressed through specific normative roles of epistemic ideas and specific actions. These roles of epistemic ideas and methodological actions have a cognitive normativity in scientific inquiry in spite of the usual controversies involved in clarifying these roles. For instance, “experimental demonstration” had normative roles among the natural philosophers in early natural philosophy, and one of these roles was to warrant empirical knowledge. One interesting epistemological trait of these roles is that the same epistemological concept could have different roles in different fields of inquiry, mainly due to the specific epistemological character of some areas of inquiry. Epistemological roles of epistemological ideas are highly contextually dependant and they are historically generated. In this sense, historical epistemology is the study of such cognitively normative roles and focuses on recognizing what is genuine knowledge, a proven fact, a good methodology, an outstanding explanation, strong evidence, and so on, in specific areas of inquiry, and how they came to be. From my point of view, the notion of “cognitively normative roles” articulates a study of epistemic normativity of rational inquiries. We can understand what exactly these cognitively normative roles are by the following idea: epistemic beliefs and methodological actions engage certain normative meanings through which they indicate, among other things, a specific epistemological status of a particular belief and/or a methodological procedure, and this indication does not just change during history, but also, and most importantly, is historically structured. Among different scientific contexts, we are able to recognize the cognitive status of epistemological ideas through the epistemic roles that they play in their particular scientific contexts. It is common to find these roles in a developed area of rational inquiry, that is, an area that has developed and articulated criteria of evidence and proof during its history. However, in some cases it could be controversial, even in developed areas, to what extent a specific epistemological idea or a particular methodological action indicates a certain cognitive status. What is less common, except in early phases of cognitive normative genesis, are debates about which epistemic status indicates a particular epistemic idea or methodological action. The important issue for a historical epistemology, as I conceive it, is that the configuration of this role is inherently historical. Contrary to developed areas of rational inquiry, in early phases of historical development, there is typically not a clear awareness of how to understand these roles. Consequently, during this time these epistemic ideas had no comprehensible cognitively normative role.



3. Cognitive normative roles embodied in inquiry practices


            If “historical epistemology” is an awkward expression, “scientific practice” is an even more problematic one mainly because, as recently Rebecca Kukla had pointed out, it often serves as a philosophical black box used to defer rather that to explain crucial epistemological problems (Kukla, 2004, p. 218). In his The Social Theory of Practices, Stephen Turner emphasized that a great diversity of things has sometimes been incorporated into the term “practices”. Turner remarks that “practice” has been used interchangeably with “tradition”, “tacit knowledge,” “paradigm,” “presuppositions”. Turner has pointed out that some conceptions of practices overlap one another and sometimes they are indistinguishable. However,  he says, “There seems to be a difference between two groups of concepts [of practices]—those that are based on the model of hidden premises of deductive theories, ‘shared presuppositions’, and those that refer to embodied knowledge, such as skills, ingrained cultural or moral dispositions, or linguistic competences” (Turner, 1994, p. 3). There is a very important problem derived from Turner’s book that faces diverse attempts to clarify the notion of practice, namely, through what kind of evidence can the elements of a practice be detected. It seems that the notion of practice has some hidden ingredients. “Hidden” means inscrutable, something that is producing effects but we have no access to the cause. We can elaborate many satisfactory explanations about the nature or the structure of such a cause, but if we have no empirical criteria for choosing among these explanations, we have no certainty about which explanation is correct even though all of them could be explicatively adequate. Unfortunately, this is the epistemological situation among the different current characterizations of “scientific practices”.

            Different authors of scientific practices try to evade the accusation that scientific practices contain hidden elements. These authors argue that their theories about scientific practices are naturalistic because they support their theories with empirical information. Nevertheless, there is an important sense in which they are not naturalistic theories, namely, if philosophy desires to be naturalistic as empirical science does, then philosophy must develop its own criteria of empirical proof for its own theories about science, not merely construct explanations in accordance with empirical information. In other words, it is one thing is to construct a theory of scientific practices using empirical evidence taken from different empirical sciences, and quite another to ensure that these theories have developed criteria for empirical proof. If there is no robust criteria to test our theories by which we pretend to explain the world, then we cannot be sure that we have empirical theories.

            A common defense to this anti-naturalistic accusation is to insist on the primacy of embodied practice as opposed to discursive practice. In this sense, scientific practices are those that refer to skills, moral dispositions, etc. According to this characterization, these practices are elaborated by knowledge embodied in skills, moral dispositions, and so on. It is an unquestionable statement that certain skills are developed for specific and concrete aims, for example, some mathematicians develop intellectual skills for solving second order equations. But is there a set of skills that can be considered exclusively scientific? Of course not, and a simple reason is that it could be the case, and certainly it frequently happens, that two scientists working together in the same practice could have a different degree of a skill or even a different kind of skill as they investigate scientific questions. when they do science. It is not possible to create an exhaustive list of skills, or moral dispositions, etc., which adequately characterizes what are specifically  (specifically?) scientific practices. And in the same line of argument, many skills that constitute a scientific practice belong to other practices that are not scientific. So, under these characterizations of practice, many different human activities could be considered as practices.

            Joseph Rouse (2002) has developed an interesting proposal about scientific practice in which he faced some of the difficulties mentioned above and developed the notion of practices as normative. As a central idea, he refuses any systematic distinction between the natural and the normative because he conceives of the distinction between natural and normative as a pernicious dualism. He took very seriously Turner’s criticisms of practices, but emphasizes that Turner, in his criticisms of the notion of practice, “fails to recognize the possibility of an alternative conception of practice whose constituent performances are appropriately regarded as answerable to norms of correctness or incorrectness” (Rouse, 2002, p. 169). Rouse correctly affirms that not all practitioners of a practice perform the same actions or have the same presuppositions, whereupon he recognizes Turner’s criticisms on practice as ‘shared presuppositions’. But, Rouse argues, practitioners and other constituents of a practice are accountable for performances or presuppositions that are inappropriate or otherwise incorrect. So the sanctions and other differential responses that would mark the incorrectness of some performances are themselves normative practices.

            For Rouse, “practices constitute objects-in-phenomena by configuring a setting within which the boundaries between one object and another make a difference, and are answerable to the boundaries they make” (Rouse, 2002, p. 289). The ideas is, as Kukla has recently written, that the natural world is inherently normative, in the sense that it is only insofar as it matters how parts of the natural world are and behave that these parts can be said to have a determinate character at all. Rouse argues, “We can understand a practice through the shared normative accountability of its constituent performances” (Rouse, 2002, p. 19), which means that various performances among a community can be held mutually accountable, without requiring any underlying regularities among them. In order to distinguish between practices as regularist and practices as normative, Rouse emphasizes a correlative distinction between two conceptions of linguistic or discursive practices. Those who identify practices with regularities and with shared beliefs or conceptual schemes, typically situate language outside the domain of practices. By contrast, Rouse argues, a normative conception of practices is best understood as incorporating a general conception of intentionality, as Brandom, Wittgenstein, and Davidson have argued. A pragmatic account of language and of intentionality more generally, considers language dynamically and normatively, that means that shared meanings or beliefs are not preexisting facts that explain the possibility of communication, but norms presumptively invoked in the course of interpreting language as communicative. 

            I agree with the idea that normativity should be central in articulating an operative idea of practices, instead of practices as regularism, tacit knowledge, tradition, presuppositions, or shared assumptions. One obvious but crucial aspect of normativity in general is its social character, in the sense that normativity normalizes and evaluates actions, beliefs, and values that conform social interaction. In science, cognitive normativity normalizes and evaluates actions, beliefs, and values associated with scientific inquiry. Thus, we can characterize a normative practice of inquiry as the social space of rational inquiry in which cognitive normativity presents clear normative effects in the actions, beliefs, and values associated with specific practices of inquiry, and the historical development of this normative practice of inquiry modifies its own normativity through time, through a feedback process. When we analyze the scientific cognitive normativity of practices of inquiry throughout history, we do not see it either as a rigid phenomenon or merely as a perpetually changing (at random) phenomena. What we notice by historical studies are the ramifications of the normativity and genesis of new practices of inquiry. (Are we noticing the ramifications/s of a genesis or just noticing the genesis? Perhaps we should discuss this point.) Late in the eighteenth century and during the nineteenth century, there were many attempts to incorporate the normativity associated with Newton’s inquiries into different areas of rational inquiry. There were attempts by Thomas Reid to normalize and evaluate empirical theories of mind through what he understood as a Newtonian cognitive normativity (obviously Reid did not use these terms). Lyell did the same in geology and Darwin in his evolutionary biology. All these attempts served to develop, after a certain time, a normative practice of inquiry with different conceptions about how to evaluate theories.  These cases were examples of the ramifications, not in all cases successful, of the cognitive normativity associated with Newton’s practice of inquiry.

            Normativity of a normative practice of inquiry does not depend on a priori principles and consequently is not a formal necessity. It is a material necessity, so to speak. Consequently, a normative practice of inquiry also changes throughout history through a feedback process of change between the practice of inquiry and its normativity. Normativity of a cognitive normative practice is historical. This idea was advanced by Rouse in the following way. He argues that among a neo positivistic tradition in the philosophy of science, “normativity is still to be grounded in necessity, but the relevant necessities are not logical or transcendental necessities knowable a priori. The world disclosed by the sciences is not merely a jumble of contingent facts, but is structured by brute, a posteriori necessities. Such necessities supposedly provide a more local, naturalistic grounding for the normativity of scientific understanding” (Rouse 2002, p. 7). Then, he says, “It is useful to begin with attempts to locate the normativity of meaning and justification in some form of social-historical necessity.” (Ibid. p. 7). In my approach to a historical epistemology, normativity is grounded not in an a priori necessity, but in an a posteriori historical necessity. As we have seen above, cognitive normativity in science is the way in which scientific inquiry usually associates certain cognitive values with specific epistemic ideas and this association is historically articulated.




[1] For example, Jurgen Renn, from the Max Planck Institute in Berlin, has recently said that, according to his research group, historical epistemology “Attempts to open up a space for exploring the relationship between all relevant dimensions of the development of scientific knowledge. Our goal comprises the reconstruction of central cognitive structures of scientific thinking, the study of the dependence of these structures on their experiential basis and on their cultural conditions, and the study of the interaction between individual thinking and institutionalized systems of knowledge. Historical epistemology in this sense requires an integration of social, cultural, and cognitive studies of science” (Renn 1996, p. 4). In this sense, historical epistemology is close to a cultural and social history of science with the idea of a “cultural system of knowledge” as one of its central notions.