Medicine Studies

, Volume 2, Issue 4, pp 245–263

Kuhnian Paradigms: On Meaning and Communication Breakdown in Medicine


    • Philosophy Department, Furness CollegeLancaster University

DOI: 10.1007/s12376-011-0063-7

Cite this article as:
Dragulinescu, S. Medicine Studies (2011) 2: 245. doi:10.1007/s12376-011-0063-7


In this paper, I enquire whether there are Kuhnian paradigms in medicine, by way of analysing a case study from the history of medicine—the discovery of the germ theory of disease in the nineteenth century. I investigate the Kuhnian aspects of this event by comparing the work of the famous school of microbiology founded by Robert Koch with a rival school, powerful in the nineteenth century, but now almost forgotten, founded by Carl Nageli. Through my case study, I show that medical science possesses some Kuhnian features. Within each school, scientists used similar exemplars and shared the same assumptions. Moreover, their research was resistant to novelty, and the results of one party were disregarded by the other. In other words, in a moderate sense, the Koch and Nageli groups worked within distinct paradigms. However, I reject the stronger Kuhnian claim that the terms used within the two paradigms were mutually unintelligible. Focusing on the semantic aspects, I argue that no account of incommensurability of reference can be given in this case, although, for sociological reasons, the two parties talked past each other. I suggest in addition that the rival scientists could have understood each other more easily if their theoretical commitments had not been so deeply ingrained, and I use the example of Pasteur to indicate that the causal account of meaning might have avoided the communication breakdown.


Germ theoryInfectious diseaseParadigmsKuhnFleckIncommensurabilityKochPasteurNageli


It is well known that according to Thomas Kuhn, the history of science is a succession of revolutions and periods of normal science. Revolutions are substitutions of old theories with drastically different new ones, while normal science periods are times in which the newly introduced theories are consolidated and extended. What is peculiar to a revolutionary change, according to Kuhn, is that the knowledge thus provided is not continuous with the previous one. A revolution does not extend or correct the previously held beliefs—it substitutes them. Thus, claims Kuhn, the traditional view of the history of science as an accumulative process can no longer be embraced. The history of science just consists of a series of incommensurable paradigms.1

The notion of paradigm designates a specific research path founded by a revolution, along which scientists are only able to recognise and work within the scope of the peculiar type of knowledge initiated by the revolutionary theory. The basic assumptions and models (be them practical or theoretical) of the revolutionary theory constitute, says Kuhn, an exemplar that guides tacitly and explicitly the work of scientists working within a paradigm. Paradigms are incommensurable if there is no possibility to decide which of the competing theories (the traditional or the revolutionary one) is closer to the truth.2

The huge influence of the exemplar—which circumscribes what problems are worth solving, the methods of enquiry, the values to evaluate improvements and the theoretical ‘core’ of any future development—ensures the unity of scientists’ work within a paradigm. Research within such a ‘disciplinary matrix’ involves a certain disregard of facts that cannot be easily explained within the current framework—the so-called ‘resistance to novelty’.3 When experience provides a sufficient number of such ‘anomalies’, a crisis emerges. After passing through a period of pre-paradigmatic uncertainties, a new paradigm is established.

Notably, one of Kuhn’s inspiring resources was Ludwik Fleck’s ‘Genesis and Development of a Scientific Fact’, a book concerned with the discovery of the Wasserman reaction (the method to detect in blood the causative germ for syphilis).4 Fleck seems to come close to the idea of incommensurability in certain passages of the book. His main goal, however, is to illustrate the resistance to novelty of scientific communities and describe the sociological means that make possible and perpetuate this kind of ‘blindness’ to contradictory evidence. Now, it is noteworthy that the evidence Fleck presents as having been ignored by the medical research into syphilis and its causes in the early twentieth century is part of a controversy born in the nineteenth century about the specificity of bacteria and their pathogenic actions. Two opposing schools were involved in that dispute—the Linaeans, whose main figure was Robert Koch, and the Unitarians (or Transformists), whose main representative was Carl Nageli.

The two schools and their research are the case study of the present paper. The main areas in which Kuhn applies his specific view on the evolution of science are physics and chemistry, and medicine has remained a field ignored by Kuhn (and philosophers of science in general). Given the influence that Fleck’s work has exerted over Kuhn’s ideas, a legitimate question we should ask is whether medicine might have paradigms in the same way in which Kuhn claimed physics and chemistry to have. Does the nineteenth century controversy about the germ theory of disease indicate the confrontation of two competing paradigms?

We should distinguish at this point between a moderate and a radical sense of a paradigm. The radical sense involves incommensurability between paradigms. The moderate sense hints to the fact that in certain periods, scientific communities share assumptions, are deeply influenced by their exemplars, resist novelty and are bound by very strong sociological connections. In other words, the moderate sense of paradigms entails that sometimes scientific communities act as if there was a relation of incommensurability between them. In this moderate vein, sociological, historical and philosophical insights could be used to show that scientists from different camps talk past each other, as if their values, meanings or perceptions were different, even though communication could be (or could have been) established.

To move from a moderate to a radical sense of paradigms is to point out that scientists from different communities cannot understand each other by any means. That is, no proper communication between rival paradigms can be established. The most important ground for such incommensurability, according to Kuhn, is the semantic one.5 Between paradigms there is a gap of meaning, which amounts to the impossibility of establishing a neutral language allowing a comparison between the claims and results of their research.

In our case study, the complex definitions and systems of classification used by the two bacteriology schools with respect to pathogenic bacteria and diseases suggest that a semantic reason for their lack of communication is possible. Does the Koch-Nageli dispute (and the competition between the Linaean and Unitarian schools) provide an example of a clash of Kuhnian paradigms in medicine? I argue in the following that at least in a moderate sense, we have in this case an example of confronting paradigms. However, I suggest that no semantic incommensurability can be convincingly shown, which entails the rival scientists involved in the discovery of the germ theory of disease could in principle have communicated with each other.

I proceed by outlining the common assumptions and exemplars within each school and pointing out the strong sociological bounds uniting them (‘Common Assumptions and Exemplars within Each School’). Then, I illustrate how each of the schools resisted novelty (‘Resistance to Novelty’). In the third section, I look at the potential reasons for semantic incommensurability and reckon in turn with two different accounts of meaning and reference—the descriptive account (‘The Description Account’) and the causal account (‘The Causal Account of Reference’)—in order to show that no gap of reference existed between the two schools. In the last section, I analyse the case of Pasteur—a French bacteriologist who was not involved in the dispute between the two German schools—and use his work in order to strengthen my claim that the Koch-Nageli dispute betrays a certain amount of unwillingness of the parties to listen to each other. I suggest that one can determine, via an empirical historical analysis, what account of reference a scientist or a group of scientists employs. Finally, I suggest that the sociological and theoretical commitments that hinder communication between moderate paradigms might be more easily overcome when the causal account to reference guides scientists’ search for meaning (‘Pasteur’s Case’).

Common Assumptions and Exemplars within Each School

What were the basic assumptions of each bacteriological school? As concerns the Unitarians, they held firm beliefs about bacteria, the role of environment and the nature of disease. One of their main claims was that bacteria cannot be classified into separate species. Carl Nageli, their most prominent representative, claimed that although bacteriologists can set out for more or less practical purposes different kinds or species of bacteria, these kinds are not in reality clear-cut and separate because in-between one can always find intermediary individuals, exhibiting gradual, almost infinitesimal differences.6 Such a lex continui governs not only in the organic realm, but also the inorganic one. One cannot establish definite border between the living and non-living; on the contrary, there is a general transition from inorganic to organic, from more and more complex organic molecules to the simplest unicellular living organisms.

Unitarians added to this picture a dynamic, evolutionary dimension. The simplest organic compounds develop from inorganic matter, if the right conditions are present. This process also occurs in the more complex stages of organic development, but importantly, not in a branching fashion, such that there are diverse, parallel lines of descendant. Although at a certain stage in the development of complex structures the morphological similarities between various organisms could suggest a branching process, their resemblance might not be due to a genetic, ‘blood relationship’, but to the same laws and environmental conditions; that is, they might in fact belong to different lines of descendant. Since morphology is most often a result of the relation to the environment, then false categorisations could well be framed on the ground of visual similarity. As Nageli wrote in 1884:

‘The overall Blutverwandschaft of the living organisms of the present day and of the whole phylogenetic system is actually nothing but a beautiful dream. But because of the unity of developmental laws, which cover the whole of the organism kingdom, it can be accepted as a general norm. Even if the organisms are not genetically connected, they behave by and large as if their true relationship to each other was a true one’7

The lex continui and the attending evolutionary ambiguity provided two (interrelated) reasons for scepticism about the ‘beautiful dream’ of a strict, definite classification of organisms. Any two organisms belonging prima facie to the same species could in fact be either the result of different lines of descendant, subjected to the same external laws and environment conditions (no ‘blood relationship’, that is), or the result of the same line of descendant, but with gradual, intermediate differences between them.

Unitarians believed that the failure to provide clear-cut classifications is especially present in the realm of bacteria, because of their less complex organisation. Being in the beginning of different lines of descendant, they are continuously created from inorganic matter and since the conditions do not strictly shape their evolution—defining, en route, their morphologic aspect—they can easily transform one into the other. Crucially, the actions of bacteria were not considered relevant to classifications. A bacterium exhibiting the effect X could exhibit the effect Y if, in the course of a few generations, the gradual, continuous transformations it had undergone allowed this. Experimentally, the Unitarians used in their research liquid media techniques for incubation, in order to prove that bacteria are not pathogenic as such, but become pathogenic in certain environmental conditions and thus can return to their harmless state. Pathogenic bacteria, if expelled from the human body, lose their harmful effect,8 and, complementary, harmless internal bacteria could transform into pathogenic ones.9 In addition, they used their laboratory findings to reject the idea of a specific organism causing a specific disease (the one species-one disease assumption, which, as we shall see, was backed up by the Linaeans) showing how bacteria associated with a certain kind of disease could undergo transformations and produce other types of diseases.10 In different liquid media, the same bacterium could exert very different effects; for example, the same germ could cause diphtheria, typhoid fever or cholera, when cultivated in liquid media. As Nageli stated in 1877:

‘If my opinion on the nature of the fussion fungi is correct, the whole species in the course of generation takes on different and changeable morphologic and physiologic forms, which in the course of decades produces sometimes souring of milk, sometimes butyric acid fermentation in sauerkraut, sometimes the ageing of wines, sometimes the putrefaction of proteins… and sometimes diphtheria, sometimes typhoid, sometimes recurrent fever, sometimes cholera. When one form of this fungus species gets into a new medium, it adapts itself progressively to the new conditions, and becomes more or less weakened, and may even lose its characteristic properties, if it does not pass directly from one patient to another’11

Germs’ natural physiology is manifested only in such liquid media. To use other media for incubation (e.g. solid cultures, as their rivals, the Linaeans, did) is to create an ‘artificial’ environment, which does not allow the observation of the real life of germs, their multiple actions and their gradual transformations.

The Unitarians controlled the Institute of Hygiene in Munich,12 and their ideas were published in an influential journal, ‘Archiv fur Hygiene’, founded in 1883. A powerful part of the sanitary movement in Germany shared their assumptions. It was represented by hygienists concerned about the conditions in which pathogenic bacteria were supposed to be created (e.g. soil inorganic matter under certain temperature conditions, and decaying and putrefying organic matter). Their aim was not primarily to stop the spread of bacteria, but to limit their creation and the possible transformation into pathogenic types. Moreover, since the pathogenic actions of bacteria were taken to be dependent on the host’s resistance, a great heed was paid to conditions that could enhance the immunity of a potential host, as economic development and good nutrition.13 Unitarian hygienists recommended street cleaning, keeping the soil free from impurities by removing putrefying nuisance, expelling of sewerage in order to avoid water contamination in the supplying area and ventilation in order to remove miasmas containing pathogenic bacteria that could arise from decaying matter or modified soil conditions.

Turning now to the Linaeans, whose main representative was Robert Koch, their main assumption was that nature exhibits separate, clear-cut species of bacteria. Every species can be characterised by its specific actions, which are constant and not influenced by media conditions. As Ferdinand Cohn (the most prominent figure of the Linaeans before Koch) was arguing against Nageli in 1875:

‘It is not true that the same germ, according to whether it grows in urine or wine, turns the one alkaline and the other ‘ropy’, or that the same bacteria form butyric acid here and transmit anthrax there, or produce a red fleck on a potato and a cause of diphtheria in a human windpipe. It is far more likely that when we have better microscopes, many organisms that now look alike will be recognised to have morphological differences upon which primary species differences can be based’14

Methodologically, all parts of the infected body should be carefully inspected to assess the presence and frequency of bacteria or their actions. The pathological germ has to be inoculated upon animals that develop similar symptoms of disease, in order establish the identity of a particular species.15 That we find a fluctuating virulence for the effects produced by certain species does not mean that the effects of germs are not constant; this simply happens when a certain colony of germs overgrows.

Misunderstandings concerning the boundaries of species, Linaeans argued, contra Unitarians, arise when cultures are not pure—which happens in the case of liquid cultures of incubation—and when microscopic analyses do not sufficiently pay attention to the morphologic differences. Only solid growing media allow the development of colonies of cells, sharing a common structure between germs; these colonies are the progeny of a single organism, and therefore, are pure.16

In 1874, Koch produced a demonstration of the character and mode of growth of the causative bacillus for anthrax, in order to prove that during a life cycle, the microorganism responsible for this disease did not change its species.17 Later, in his 1878 paper on wound infections (already translated in English in 1880), Koch developed the idea of a specific organism causing a specific disease.18

‘A distinct bacteric form corresponds, as we have seen, to each disease, and this form always remains the same, however often the disease is transmitted from one animal to another. Further, when we succeed in reproducing the same disease de novo by the injection of putrid substances, only the same bacteric form occurs which was before found to be specific for that disease.’19

It was an assumption, a starting point that Koch set to follow, against theorists like Nageli who did not see ‘any absolute necessity for dividing them even into two distinct species’20; Koch was adamant that ‘till the cultivation of bacteria from spore to spore shows that I am wrong, I shall look on pathogenic bacteria as consisting of different species.’21

This assumption of specificity led him further to the development of experimental procedures meant to identify a certain species as the cause of a specific disease.22 The formulation of the respective experimental ‘postulates’ varied in different papers published by Koch between 1878 and 1884, but its practical core—entailing that specific microorganism should be present in all cases of a disease, should be obtained in pure cultures outside the host and should produce the same symptoms, when inoculated into another susceptible host—came to permeate the research of his group of scientists23 and underlined the reported discovery of bacterial causes for tuberculosis (1882), cholera (1884), diphtheria (1884), typhoid fever (1884), gonorrhoea (1885) and pneumonia (1886).

The Linaeans formed a well-defined social body that became more and more influential following Koch’s appointment to the Imperial Bureau of Health. A powerful journal, ‘Zeitschrift fur Hygiene und Infektionskrankheit’, was used to publicise the research initiated by Koch and his colleagues. They controlled the Prussian Privy Council, the Cholera Commission, and a large number of institutes of hygiene. The Linaean group of hygienists emphasised the substantial importance of the invading germ over host resistance, claiming that the presence of a real pathogenic bacterium implies, regardless of the defence system of the host, the emergence of a disease.24 Hence, they aimed at more strict and active measures in order to stop primarily the spread of bacteria, insisting on border controls, quarantines, boiled water, strong state interventions and health registration on a national scale.

Resistance to Novelty

Scientists in both schools proved themselves reluctant to accept evidence that did not fit in with their theoretical beliefs. With respect to the Linaeans, the dispute over tuberculin is a good example of their ‘resistance to novelty’. Tuberculin was the ‘miraculous’ medicine designed by Koch to cure tuberculosis, the most devastating disease of his age. Koch kept its composition secret, revealing only that it was ‘a brownish, transparent liquid, which does not require special care to prevent decomposition.’25 Later on it was revealed that it came from tuberculosis pure cultures dissolved in glycerine (and that Koch did not know precisely its components).26 The starting point for devising it was that in laboratory-infected animals, the lesions produced in lungs were initially very active, with an abundant presence of Bacilli tuberculosis. Then, noticed Koch, along with the disappearance of bacilli in the affected sites, the lesions become necrotic and seemingly non-progressive. Tuberculin injections, therefore, were supposed to hasten the stages of disease, in order to reach more quickly the necrotic, non-progressive state and thus to avoid general, systemic damage.27 At the same time, tuberculin produced an epidermal local reaction and general fever in a diseased person, and these manifestations were taken as diagnostic signs in doubtful cases.28

Announced with great enthusiasm in 1890, this ‘miraculous’ medicine was immediately accepted at a time when Koch had become a figure of ‘national pride’. The way in which the Linnaeans evaluated tuberculin’s ambiguous effects the following period is, however, a ‘pathognomonical’ case of resistance to novelty. We should at this point return briefly to Ludwik Fleck and his description of the behaviour of scientific communities that inspired Kuhn so much. Fleck describes five stages of the kind of reaction to evidence specific to novelty-resistant scientific communities: (1) A contradiction to the system appears unthinkable; (2) What does not fit into the system remains unseen; (3) Alternatively, if it is noticed, either is kept secret, or; (4) Laborious efforts are made to explain an exception in terms that do not contradict the system; and (5) Despite the legitimate claims of contradictory views, one tends to see, describe or even illustrate those circumstances that corroborate current views and thereby give them substance.29

These stages are clearly exemplified in the case of Linaeans’ tuberculin. For one thing, what remained for a good period of time unseen by Koch and his proponents was that actually healthy persons could exhibit the high fever manifestation when injected with tuberculin. Rather late, the explanation was given, in terms of the allergic, immunological hypersensitivity to germs enclosed in tuberculin.30 For another, persons whose diagnostic dermal test with tuberculin was positive could just present a primary tubercle (a unique granulomatous formation, representing the reaction of tissues to the Bacilli tuberculosis infiltration, localised at the apex of the lung, and not affecting the general state of the organism),31 which was not a stage of disease. After an experiment in which he had injected twenty healthy individuals and had observed the high fever reaction in some of them, Erich Peiper, a well-known proponent of Koch ideas, concluded that those were cases of ‘latent tuberculosis’, although his observations were a salient contradiction to Linaeans’ expectations.32

Moreover, the necrotic lesions were not a sign of a non-progressive state of the disease. After the death of tuberculosis patients who had been injected with tuberculin, it was claimed that they had been too much weakened by the disease to react fully to the drug. In the same vein, misinterpreted autopsies were used to back up the claim that tuberculin leads to a standstill point of the disease, by emphasising the necrotic lesions described by Koch. Even after Virchow—who had shown that on the fringes of necrotic lesions were still present fresh tubercles33—raised anatomical counter-arguments, Koch’s proponents did not give up. They continued to stick to their initial theory, claiming that there were types of tuberculosis responsive to treatment, while others were not. Koch himself was unable to accept the fundamental failure of tuberculin as a medicine all the rest of his life.34

On the side of the Unitarians, Max von Pettenkoffer, the head of the Unitarian school of sanitation, offered the most salient proof for resistance to novelty. In the 1892 Hamburg cholera epidemic, the measures taken by the officials were initially devised by von Pettenkoffer, in line with the sanitary approach of the Unitaritans. Notably, the water supply problem was not taken into account because pathogenic bacteria were supposed to come in contact with the soil in order to generate the cholera poison. Although epidemiological data had clearly suggested that the epidemic was much less severe in a nearby town in which the water supply (a river) was filtrated through a sand bank, the water problem was not revisited until the Prussian central government decided to impose upon the town the severe measures of the Koch-influenced school of epidemiology.35

The results were dramatically effective but that did not convince von Pettenkoffer. In 1892, von Pettenkoffer and his assistant drank water containing live cholera germs in order to demonstrate that the measures taken by Linaeans (e.g. boiling water) were useless.36 That is, he continued to back up Nageli’s argument that pathogenic bacteria, once expelled from the body of a sick person, become harmless; and that, on the other hand, pathogenicity has to be construed in relation to a certain exterior factor (like the ‘modified soil’), acting upon bacteria and transforming them in order to produce the disease. Of course, in this vein, living cholera germs, taken directly from experimental cultures (just like the ones from a water supply not contaminated by sewerage), were not supposed to cause any harm. Fortunately, von Pettenkoffer contracted a very mild form of the disease; his assistant fell more seriously ill but recovered after a few days since the liquid samples that the Linaeans had charitably prepared for them were low in bacilli.37

Of course, one should bear in mind the argument that in general, Koch’s conception of infectious diseases was too restrictive (by placing the entire load of pathogeny on bacteria and viewing human organisms as defenceless), and that von Pettenkoffer’s theory, with its multi-level construal of causation, was unduly disfavoured for decades after his lost confrontation with Koch.38 But the case of the Hamburg cholera epidemics with its water filtration episode is a clear case of resistance to novelty on the part of the Unitarians, just like the case of the tuberculin is such a case on the side of the Linaeans.

Semantic Incommensurability?

In his late writings, Kuhn places an emphasis on the semantic grounds for incommensurability. While the full scope of his views on semantics is still a matter of debate,39 what is certain is that the main reason why scientists from rival camps should fail to communicate is that the reference of their key terms might change, with the move from one paradigm to the other.40 In contemporary discussions, two main views on meaning have taken central stage—the description account and the causal account of reference. Kuhn was a proponent of the description account, but in the following two subsections, I shall also discuss the claim of semantic incommensurability (for our case study) with respect to the causal account (and also with respect to a mixed view that combines descriptive and causal elements).

The Description Account

According to this account—based on Russell’s analysis of descriptions/proper terms and Frege’s differentiation between the sense and reference of an expression—what makes it possible for terms to hook onto the (extra-linguistic) world is an intensional activity of speakers, which connects linguistic items to a reference/extension that satisfies certain descriptions associated with the term.41 Thus, this view assumes, on the one hand, that knowing the meaning of a term is just a matter of being in a certain psychological state and hence being aware of its intension/description. On the other hand, it assumes that sameness of intension/description entails sameness of extension.42 Insofar as science and its language are concerned, it follows, on this account, that the reference of scientific terms is established by descriptions depending upon the theories these terms are embedded in. Moreover, that a certain term may change its meaning (across paradigms) should be a consequence of the dissimilarity between the theories held by rival scientific communities. This dissimilarity should prompt different descriptions and hence difference references for terms across paradigms.43

It is useful to distinguish between two ways of viewing the range of the dependence of descriptions on the respective theories—namely the thick and thin intensionalist views. To start with thick intensionalism, this view says that the theory dependence of the intension is very broad, in that the whole or most part of the theory is involved in the intension/description of any scientific term. Notably, this view is embraced by Kuhn,44 and its consequences on semantic incommensurability can be put as follows: If T1 and T2 are two theories belonging to rival paradigms and ‘t’ is a term used by both groups of scientists, the shift of the meaning of t, with the move from T1 to T2, may entail a shift of the reference of the term t. The descriptions offered by the two intensions, thickly dependent on T1 and T2, respectively, are different.

As an aside, we should note that Kuhn’s thick intensionalism was used as an explanation of scientists’ resistance to novelty and, relatedly, for the need of revolutions to overthrow well-established paradigms. Indeed, if the meaning of a term depended on most of the involved theory, scientists could not move from a paradigm to another step-by-step, but only through a revolution. On the other hand, within a paradigm, scientists resist novelty—even if it might contradict only a part of their assumptions—because otherwise they face a meaning shift for all their terms. Moreover, a thick approach of intensionalism was consistent with the holistic way in which Kuhn thought that scientists within a paradigm acquire concepts. Instead of getting hold of them via straightforward definitions, maintained Kuhn, scientists learn the meaning of their terms via their use in exemplary situations. In order to understand these exemplary situations, they have to acquire the whole set of beliefs about the world contained in their paradigm.

Now, one way to question Kuhn’s thesis of meaning variance is to narrow down the dependence of terms on the theory, such that their meaning is dependent on only a part of the theoretical assumptions. Instead of a thick intensionalism, we would thus have the thin intensionalism mentioned below. Another strategy is to consider the reference of theoretical terms as being fixed independently of the theory, as in causal theory of reference, which will be analysed in the next subsection. In the present subsection, I will stick with the descriptive account and see whether the thin intensionalist approach can offer an alternative to semantic incommensurability in our bacteriology case study.

The thrust of the thin approach to the intensions of theoretical terms is that the reference of key scientific terms might remain unchanged, even in the case of a substantial difference in the theories advanced by rival scientific communities. Paradigms, we should note, do have common assumptions—otherwise they could not be considered rival.45 If those assumptions were sufficient to fix the reference via associated descriptions, the theory change would not influence the meaning of key terms.46 That is, if the description of a term ‘t’ is given by an intension dependent on the T-core (the part of theoretical assumptions shared by T1 and T2), the term will have a fix reference in spite of any previous or further disagreement between rival paradigms.

Now, the key notions in both bacteriology (candidate) paradigms of our case study were infectious disease and (pathogenicity-involved) bacteria.47 To recall briefly Ludwik Fleck’s seminal work mentioned in the Introduction to this paper, one of his claims from ‘The Genesis…’ was that the meaning of syphilis—the disease supposed to be detected by the Wasserman reaction—changed several times from the sixteenth century onwards, in close connection to more or less scientific theories employed in order to explain or cure the disease. These explanations ranged the astrological constellation, religious teachings or the curative effect of Hg to germ theories of disease.48 Hence, the question we should ask is as follows: Was there a difference in the meaning of various diseases, as concerns the Unitarians and the Linaeans? It is important to note, on a methodological level, that both the thick and thin intensionalist approaches should be considered when investigating the meaning of a term or concept, in order to avoid the charge of arbitrariness.

In principle, the concept of a contagious disease might be influenced by symptoms, causes, means of contagion and treatments (or sanitary measures).49 As concerns symptoms, in the second half of the nineteenth century, physicians generally accepted for each disease a set of more or less pathognomonic signs, expressed in an observational vocabulary that—though not specific to any of the two bacteriology schools—was used by them extensively. The main area of disagreement had to do with causes and treatments (including sanitary measures) and even there, on a very general level, they shared some common points in their explanations of pathology and prevention. For instance, both schools took germs as being causally involved in contagious pathology (and not miasmas), and their preventative treatments converged on some outcomes (e.g. the removal of sewerage).

Now, if the descriptions associated with the notion of disease and any particular disease terms involved these general assumptions, on a thin intensionalist approach, a shift in reference across the two communities would not be tenable. However, if the descriptions in question were dependent on the whole (or most part of) theories involved, as the thick intensionalist approach says, the ravaging disagreement about the specific nature of pathogenic bacteria between Lineans and Unitarians could have led a variance of reference. Since in the latter, Kuhnian scenario, the meaning of the disease concept is influenced by the meaning of (pathogenicity-involved) bacteria, we should look in the first instance at the causative agents of diseases and then draw the implications over the meaning of disease notions, as such. An illustrative example of a disease is cholera, whose causes were the source of one of the most important disputes in that period, given the gravity of cholera epidemics. The Linaeans had identified Vibrio cholerae as the germ involved in the disease and, of course, claimed it was a well-defined, specifically acting species. Unitarians’ reply was in line with their specific assumptions—they held that Vibrio cholerae, as a well-defined, specifically acting species, is indistinguishable from other vibrios.50

On a thick intensionalist approach, the intension associated with the causative agent for cholera, as conceived by Koch’s group, would have to describe it as a separate, fixed species, state the causal contribution in the pathological process and involve the aetiology algorithm developed by Koch: ‘Vibrio K.—a bacterium with the morphological traits of the genus vibrio found in the intestines or blood of patients who present cholera symptoms and which, after having been injected into a healthy animal, is again isolated in the intestines or blood of the animal exhibiting symptoms of cholera’. On the other hand, we should take an intension of the causative agent of cholera, according to Nageli’s group, to describe its various actions and its possible transformation into another a member of a different species: ‘Vibrio N.—a bacterium sharing morphological aspects of genus cholera, which is isolated in the intestines or blood of patients who present symptoms of cholera and which, after having been injected into a healthy animal, can again be isolated from the intestines or blood of the animal presenting symptoms of dysentery.’

These two intensions share a part of the description but are, on the whole, different. Does this lead to a shift in reference? Kuhn thought that this is what happens in rival paradigms. But a crucial response to his claim, formulated by W. Newton-Smith and more recently, Alexander Bird, is that if an intension gets too thick and complex, the danger is a reference failure.51 When the description depends on the whole theory, the term will refer only if the whole theory is true. When the description depends on a part of the theory, the term can still refer even if some parts of the theory are mistaken. The more complex an intension is, the lower the chance that anything would fit it. It is easy to see that from the point of view of contemporary medicine, no entity could satisfy the description of vibrio N. The reference in question is empty. Furthermore, no matter how we varied the thick description that could have been associated by the Unitarians to the term Vibrio cholerae, if that description involved the false beliefs that Nageli et al. held about the nature of germs, the correspondent reference would ineluctably be empty. Thus, thick intensionalism cannot make room for semantic incommensurability. On the other hand, if we considered the reference of the germ in question as fixed only by the common part of the above-mentioned descriptions, the Unitarian and Linaean reference would be identical. That is, vibrio N. and vibrio K. would describe a bacterium with morphological aspects of the genus ‘vibrio’, found in the intestines of patients presenting symptoms of cholera. Thin intensionalism also rules out semantic incommensurability on this level.

We can return now to cholera as a disease, as such, and what we see is that—after considering the thin and thick intensionalist variants—the same conclusion follows. We could have different descriptions associated with the term ‘cholera’, if its intension was dependent, in a thick manner, on its causative agent, but we could not have a genuine shift in successful reference. The thick description of cholera N. would be dependent on a non-referring ‘Vibrio cholerae’ term, which would also produce the lack of reference for the disease term, as such. Cholera N. defined using cholera symptoms and the presence in the blood of vibrio N. is just an empty word. On the other hand, a thin intensionalism would keep the same reference for vibrio N. and vibrio K. and, accordingly, the intensions of diseases, cholera N. and cholera K., respectively, would be the same. No semantic incommensurability could be proven. Again, the same rationale can be applied to other points of disagreement between the two schools, e.g., Unitarians’ claim that pure culture of bacteria have to suffer the action of certain soil condition in order to become pathogenic, and the laboratory solid media are ‘unnatural’ environments for germs, opposed to the Linaean assumption that solid media of culture are indispensible for grasping the bacteria effects.

We see that on the description account, no incommensurability of reference can be indicated. Does the Kuhnian case fare any better when the causal account of meaning is reckoned with?

The Causal Account of Reference

I have noted in ‘The Description Account’ that according to the description account, what makes it possible for terms to hook onto the (extra-linguistic) world is the intensional activity of speakers, which connects linguistic items to a reference/extension that satisfies certain descriptions associated with terms. This view assumes, in other words, that knowing the meaning of a term is just a matter of being in a certain psychological state (and hence being aware of its intension/description) and that sameness of intension/description entails sameness of extension. These assumptions were famously challenged by Putnam by means of a thought experiment intended to show that the beliefs held by a speaker about the natural kind water before the discovery of water’s chemical formula could not have fixed the reference of the corresponding kind term ‘water’. I will briefly espouse in the following Putnam’s rationale, since the causal account Putnam founded represents a crucially important answer that could be raised against the Kuhnian challenge of incommensurability and, as we shall see in ‘Pasteur’s Case’, might play a significant explanatory role in disentangling the reasons for the sociological communication breakdown occurring sometimes between scientists.

Putnam’s strategy is to show that even if a speaker held certain beliefs about water before 1790 (the date of the discovery of water’s formula), his/her mental states would not have been sufficient to fix the reference of this term, since similar mental states (concerned with the superficial properties of water) could have been possessed by a (‘identical’) speaker, with respect to a natural kind that was characterised by different microstructural properties and that yet populated the lakes, rivers, oceans, etc. The latter speaker would have said something false when pointing at that liquid from rivers, lakes, oceans, etc. and uttering ‘This is water’. What this shows, claims Putnam, is that the reference of the natural kind term ‘water’ cannot be fixed by the intensional activity of speakers, i.e., cannot be adequately accounted for by considering only the mental states/beliefs of speakers, per se.52 Natural kind terms successfully refer, Putnam argues instead, in virtue of an indexical activity of the speakers, who stand in appropriate causal relations to the natural kinds in question. These causal relations cannot be captured by any description or at least any description that does not contain indexical terms. ‘Meanings just ain’t in the head’ declares Putnam, where, again, this slogan should convey that the reference cannot be fixed by the mind alone, in the absence of appropriate causal relations to the (non-linguistic) reality.53

Putnam also offers an explanation as to how precisely the reference is picked out in this causal way, given that it seems implausible that each speaker could perform such ostensive/indexical acts for all their (kind) terms. Speakers, according to Putnam, baptise in the first instance paradigmatic samples of the natural kinds in question. Then, the terms are introduced into the usage of linguistic communities. Since in the act of dubbing, a causal relation is established between a certain term and its reference (natural kind) indicated indexically, the term at hand has in the linguistic community where it is introduced a ‘tenacious’ linguistic behaviour, in the sense that its extension is always the kind whose paradigmatic sample was involved in the dubbing, irrespective of what more or less appropriate descriptions one speaker or another might associate to the term..54 All the instances of the natural kind in question will be part of the extension of the term because of an indexical same relation with the paradigmatic sample involved in the dubbing, a relation that ‘[it] may take an indeterminate amount of scientific investigation to determine’.55 In science, the act of dubbing might be performed slightly different from everyday life, because scientists often do not have direct (observable) access to the natural kinds they study. However, in the absence of an (observable) sample, postulating an entity that is causally connected to an observable phenomenon is sufficient for the dubbing.56 Thus, in science as anywhere else, the indexical activity of the speakers who introduce terms into usage is decisive for the meaning of these terms.

Insofar as the Kuhnian charge of incommensurability is concerned, what follows is that, on a causal account of meaning, the difference in theory between rival scientific paradigms could not crucially determine the fixing of the reference. Although scientists from rival scientific communities could have associated different intensions with one term or another, the fact that they intended to explain some ostensible effects, by dubbing postulated entities, entails that their reference must have been the same.

Exactly this point can be made about our case study, and accordingly this account of meaning seems to offer another plausible solution to the problem of reference shift as concerns the German bacteriologists. On the causal account, in spite of all their theoretical disagreement, their reference seems to have been the same, because it was either established by dubbing samples of bacteria (say, microscopical slices) or postulating bacteria in order to explain ostensible, specified effects (symptoms of diseases). Take as an example the controversy between Koch and Nageli’s schools on the causative germ of the anthrax symptoms.57 Bearing in mind Putnam’s famous thought experiment, we can imagine two almost identical planets, Earth (E) and Twin Earth (TE), in which speakers share the same beliefs. On both planets, anthrax has a certain set of symptoms, and in the blood of anthrax patients can be identified a germ with morphological traits of genus bacillus. On E, bacillus anthracis is a separate, well-defined species, and the loss of its pathogenicity does not involve a transformation of its essential properties; more precisely, it remains the same species after the virulence was reduced (bacillus anthracis K.). On TE, bacillus anthracis undergoes a transformation along with its loss of virulence, that is, it transforms into another species (bacillus anthracis N). In all the other aspects, the two planets are the same.

On E and TE, we have two observers, Observer 1 (O1) and Observer 2 (O2), respectively. They do not know much microbiology but share the belief that Bacillus anthracis is the cause of anthrax. The two observers are just able to recognise morphological traits of genus bacillus (these are superficial traits of species on a microscopic examination) and anthrax symptoms.58 If the extension of the term ‘Bacillus anthracis’ on E was the same with its extension on TE, O2 on TE would say something false when pointing at a microscopical slice and saying ‘this is bacillus anthracis’. But since the experiment is set up symmetrically between E and TE, there is no reason to accept that people on TE are wrong and the ones from E are right. The Putnamian conclusion would be that O2 on TE, in using the term ‘Bacillus anthracis’ is referring to the bacillus that undergoes a species change, while O1 on E is referring to the bacillus that belongs to a stable species. Even though the beliefs they possess are the same, their referents are grounded in the causal connection between the use of their terms and each of the bacilli. What this Putnamian scenario tells us is that in our case study, when it comes to the fixing of the reference for controversial notions like bacteria and diseases, the observers’ bacteriological beliefs do not play a part (or at least do not play the decisive part).

When analysing the dispute between the two rival paradigms over the causative germ for anthrax, as well as when analysing any other dispute between the two schools, we might not even have to take into account their beliefs and arguments, different as they may be. We just have to suppose that there was the same causal mechanism grounding their reference in connection to the same bacilli and anthrax symptoms. The causal account of reference is not a panacea, and surely, there are certain problems associated with it. But these problems are not unsolvable, and they do not pose threats for our line of argumentation. Let me exemplify. One of the most important difficulties faced by this account is that simply dubbing, in an indexical way, a sample or an entity causally connected to certain ostensible phenomena can be insufficient to establish successful referring. The causal link between a term and the reality should relate the term only to its reference. Scientists use natural kind terms, and the samples or the postulated entities can be conceived, at the same time, as part of different natural kinds. Hence the question—which of the possible natural kinds is picking out a user when dubbing a sample? For instance, if someone simply points to a dove and says ‘Let’s call this M.’, does he baptise a bird, a vertebrate or an animal? Or, does he baptise a spatial or temporal part of the dove? Known as the qua-problem, this weakness of the causal account has prompted some to require that the causal fixing of reference should be attended by a description. Such a description should allow the use of a term causally connected to a determinate reference. But could the intrusion of description shatter the case made by the causal account against Kuhnian incommensurability in our case study?

The answer is no. Faced with the qua-problem, we should determine what descriptions the two groups of scientists could have associated with the causal links to their germs. The problem of reference shift seems to arise again, because these associated descriptions can be different, but the distinction between thick and thin descriptions is again useful. Thus, a sufficiently general description (a thin description) would preserve the advantages of the causal account of reference. The entity that is causally responsible for the anthrax symptoms would have as an associated description accepted by both parties ‘a bacterium with the morphological traits of genus bacillus’.

The more problematic case would occur if the associated descriptions were more complex, involving more points of disagreement (whether bacteria can or cannot be grown on a solid culture, whether one can or cannot cause other symptoms in another host organism, etc.) In the case of the ‘pure’ description account, we have concluded that these thick descriptions may very well lead to a reference failure. They end up picking out nothing. However, such a conclusion seems odd in this case, because we still have the causal connection between the use of the term ‘bacillus anthracis’ and its reference. It would sound equally odd to say that scientists’ terms can successfully refer, although they associate to the causal link false, thick descriptions. If O1 on E associated a Unitarian description to his causally fixed reference, would the reference of his ‘Bacillus anthracis’ still be a discrete species or his reference would be empty? Or perhaps he/she would have a different reference?

At this point, it is worth recalling that the causal account of reference is mainly concerned with the initial step in which speakers use terms, establishing their reference. When scientific terms are involved, it could well be misleading to focus on the use of terms at a later stage of scientists’ research because, certainly, the descriptions they could have associated with a term grow more and more complex and the baptism most probably has been already made. In other words, the dubbing and the possible associated descriptions should be searched for in the context of the discovery of a certain entity, and it seems like the problematic case of thick, false descriptions attending the dubbing is not involved in this discovery context. Of course, it is still a very interesting question what happens with the reference when the causal relations are attended by thick descriptions. But again, in the context of the discovery of a certain entity, scientists’ knowledge about what they baptise is not a very complex, thick one and accordingly, it is very plausible that the associated descriptions would be thin, less complex ones.

To take stock, it seems fair to say that on both the causal and description account, there is no room for any shift of reference and hence for Kuhnian semantic incommensurability between the two scientific communities. The two schools of our case study did not represent paradigms in a radical sense. They were, however, moderate paradigms and we are still left with one explanatory task. Why did the scientists in the two schools disregard each other and did not consider the evidence that contradicted their assumptions? One answer, of course, would be to appeal to the sociological context of each school—they represented entire scientific communities, with (social, medical and even political) commitments and a well-integrated place in the academic arena of the German medical research; what seems particularly interesting in this regard is that that Koch and his school of bacteriology and public health were employing military metaphors derived from the (internal and external) German politics at the time and were backed up by the conservative, central government, whereas the views of Nageli and the sanitary movement represented by von Pettenkoffer were consonant with the liberal stances stressing free circulation, commerce and improvement of life conditions.59

This sort of sociological/political answer, popularised by authors such as Bruno Latour in all areas of science, seems plausible, in our case at least.60 But besides that our analysis of the semantic incommensurability claim can help us to advance in the explanation of their communication breakdown. I will show in the next section that their capacity to communicate and pay heed to their rivals’ results depended on how deep their theoretical commitments were, not just in a sociological sense but also from the point of view of the account of meaning they presumably employed. The case of Pasteur, a French bacteriologist who was able to distance himself from the disputes in Germany and pursue his own research, will be central to my argument.

Pasteur’s Case

The point I tried to bring home in the previous section was that no matter what account of meaning was actually employed by the two groups of scientists, no semantic incommensurability can be proven. To stress, I did not assume that one or the other account of reference was definitely used by the bacteriologists in question. When trying to explain the reasons for the communication breakdown, however, we should ask ourselves how they (took their terms to) refer to the world. Of course, in the philosophy of language literature, proponents of the description account often claim that the causal account could not work, whereas proponents of the causal account affirm the same about the description account. But there is no consensus on this matter and, as I have shown in the presentation of Putnam’s views, his slogan does not entail that descriptions could not play a role in reference fixing.61 Indeed, descriptions are actually needed when the qua-problem shows up and in the end, the distinction between the causal and descriptive positions might well be a matter of degree. Hence trying to see—by way of a historical, empirical analysis as it were—where the balance inclined in bacteriologists work (i.e. towards the description or causal side) is an interesting enterprise, which, as I have mentioned, might play an explanatory role in the communication breakdown dilemma.

My starting point is that by studying how discoveries were made in bacteriologists’ research, one could see whether the reference of pathogenic bacteria was fixed in a descriptive, intensionalist way or a baptismal, causal one. In the former case, the discovery of a significant scientific fact should have been announced by the conceptual framing of the discovery (or at least, these two aspects of enquiry, the conceptual and experimental one, should have been simultaneous). In the latter case, the situation should have been reversed, in that the experimental novelty should have been firstly acknowledged and the conceptual explanation should have come afterwards.

Kuhn contends that the former case is the rule,62 and we can argue that for the most part, in Germany, germ research typified Kuhn’s view. Most of researchers working within Nageli and Koch’s group were either physicians or botanists. Accordingly, their theoretical assumptions were very intricate and complex. They initially learnt a complicated system of classification of bacteria before starting to study their relationship to diseases, and, accordingly, the acquired morphology and physiology were already grounding these relations in a certain sense. That is why their discoveries were somehow ‘expected’. However, in France, the latter case seems to be correct. Pasteur was a chemist and his beliefs were not so deeply grounded and articulated as the German scientists’ were. Pasteur and his colleagues were committed to more practical outcomes and used to say that they are not interested in theories, but only in results.63 It is even possible that until 1882, when the proponents of Koch accused him of being a Unitarian, he had not even heard about Nageli.

Pasteur’s conclusions were partly similar to Koch’s conclusions, partly on a par with Nageli’s ones. Without having been initially committed to any of theories at stake in Germany, just on experimental grounds, Pasteur rejected the Unitarian-like claim about spontaneous generation64 and approved the Linaean view about separate, discrete species. However, Pasteur admitted that the virulence of germs can increase after successive inoculations, and that there are forms of germs bearing a much reduced pathogenicity.65 Significantly, all these conclusions were drawn using a less complicated system of classification of bacteria, more useful in practice. The scientists who were part of the ‘naturalist tradition’ were naturally unhappy. For instance, Ferdinand Cohn, a great botanist belonging to the Linaeans group, complained about the loose manner in which Pasteur used with ‘sovereign arbitrariness’ names of bacteria, ‘above the rules of botanical nomenclature.’

‘He (Pasteur) speaks sometimes of vegetarian cryptogames microscopiques, sometimes of animalecules, or of champignous or infusoires, and without any proper differentiation calls the same things toraculacees, bacteries, vibrioniens, monades, which elsewhere are called Mycoderma, Mucor, mucidinees or yeasts (levures). With sovereign arbitrariness, Pasteur sets himself above the rules of botanical nomenclature’66

Was Pasteur working in an ‘unscientific’ manner? It is hard to say but there is something else here, beside the indignation that Linaeans or Unitarians expressed about his methods. To be more precise, I think that what seemed to German scientists a ‘sovereign arbitrariness’ was in fact a method to classify germs—using simple morphological traits and disease symptoms with which they were associated—such that the reference of the scientific terms employed in his classifications was causally connected to disease symptoms. In other words, I suggest that Pasteur exemplified the causal account of meaning. Let us look at one of his discoveries to see why this might have been the case.

Pasteur devised in 1881 the vaccine for anthrax, using a form of Bacillus anthracis with a reduced virulence, obtained in laboratory. This vaccine was not very well welcomed by the Linaeans, although Pasteur had repeatedly said that the Bacillus remains the same even after its virulence was reduced.67 On the side of Unitarians, there was a greater enthusiasm, because these less violent forms of Bacillus anthracis reminded them of Hans Buchner’s experiments in 1880 (Buchner contended that he had proven the disappearance of virulence from the anthrax bacillus on a serial culture manoeuvre because a new, intermediate species emerged, located somewhere between Bacillus anthracis and Bacillus subtilis—the hay bacillus).68 Pasteur’s study of anthrax started from Casimir Davaine’s identification of a bacillus in the blood of sick cattle in 1963,69 and Koch’s own work on the life cycle of bacillus in 1874, but also from the perplexing results reported by Paul Bert in 1877, according to which anthrax symptoms can be induced by inoculation using samples of blood in which the bacilli had been destroyed by oxygenation.70 In order to verify the causal relation, Pasteur used a little amount of blood from sick cattle, growing it in a liquid culture. The medium in that liquid culture allowed a very rapid multiplication of bacilli. A part of the multiplied germs were used in another, more diluted culture and so on, until about a hundred cultures of germs were grown. Each of them proved to be capable of causing anthrax symptoms in a testing animal. This was his famous method of serial cultures.71 Let us give a closer look at this procedure. The starting point was a sample, a sample of blood from an animal presenting anthrax symptoms. The assessment that indeed, the cause of symptoms was related to that bacillus, was to grow successive cultures originating from the sample. It is precisely this aspect that hints at the fact that Pasteur used the causal theory of reference.

In a seminal paper discussing the discovery of the formula of water,72 Alexander Bird points out that chemists could have made two assumptions about the reference of their research. The former was that H2O is the chemical formula of water, and nothing other than H2O could have the surface properties of water. The latter was that H2O is the formula of that single substance of which all the actual and potential water samples are instances. The former assumption, says Bird, is closer to a description account of their reference, because it uses the surface properties of water as a description that fixes the reference. The latter assumption is closer to a causal account because, implicitly, it does not exclude that for instance, on Mars there could be a substance with the surface properties of water. It just refers to that water, of which scientists have samples or could potentially obtain samples (rivers, rain, etc.). The chemists involved in the discovery of the chemical formula of water, according to Bird, used the latter assumption and did not even take into account the former one.73

Bird’s analysis can be quite illuminating in our case because we can see that Pasteur’s assumption was analogous to the latter assumption made by Berzelius et al. After having studied the effects of bacilli contained in the blood of anthrax cattle, of which all the actual and potential samples of blood from sick animals would be instances, Pasteur came to admit that the bacillus was the cause of symptoms. On the contrary, in his research, Koch used an assumption analogous to the former one. When, as a young doctor, Koch met for the first time in Breslau Ferdinand Cohn (who was at that time the head of the Linaean school of botany), and presented him his studies on anthrax, his whole demonstration of the life cycle of the microbe aimed at proving that the causative germ for anthrax was precisely that bacillus (as described and understood by Linaeans) and not any other possible germ, apparently sharing the morphology of bacillus (as Unitarians had claimed).74

Pasteur observed, by chance, that after using in inoculations cultures passing through generational cycles and object to oxygenation, the severity of the resulting symptoms reduces. The conclusion (first drawn with regard to chicken cholera and then used for anthrax and the anthrax vaccine),75 which one can obtain pathogenic germs with a reduced pathogenicity, was drawn easily, because the reference of b. anthracis had been established using the same culture samples, in causal connection to the same symptoms. That sort of conclusion would have never been drawn by Koch, because his bacilli had been defined, previously, using the complex and overloaded intension of Lineaeans. That Linaean description did not allow any kind of change in the rigid one species-one disease relation. In other words, in the situation of the decreased intensity of anthrax symptoms, the Linaean thick description was no longer synonymous to b. anthracis, whereas Pasteur’s rigidified reference was still working. Indeed, Koch’s reaction to Pasteur’s publications of 1880–1881 was to claim that the ‘attenuated’ cultures were in fact merely contaminated with common non-pathogenic organisms.76

We can thus justifiably claim that Pasteur was more committed to a causal account of reference than to a description account, whereas the two German schools analysed in this paper seem to have been more committed to a description account of reference. Of course, that does not mean the causal account could not have characterised some of the Linaean and Unitarian references. After all, they also used the symptoms of diseases as an important aid in their attempt to classify bacteria as accurately as possible, and it might have been that baptisms have occurred during their laboratory research. As I have remarked in the beginning of this section, the causal and the description account do not exclude each other. What I wanted to suggest was that they were closer to the description account. If so, we can further ask whether this was a thick or a thin description account. Pasteur is helpful in this regard.

Pasteur could not have endorsed some of Linaeans’ conclusions if they had not shared the same object of research. In other words, in those disputes in which Pasteur was also involved, no matter how (causally or by description) the reference was fixed for each of them, it was a common one. Similarly, he could not have agreed on some Unitarians-like conclusions if their reference (no matter how it was fixed) had not been the same. So it seems very plausible that at least in these aspects of bacterial controversies in which Pasteur was involved, a thin intensionalism (the same reference) characterised German bacteriologists’ research.

Pasteur’s figure should not be idealised. As Geison’s recent, unadorned biography has shown, Pasteur owed his success not just to his scientific acumen but also to his polemical and diplomatic flair due to which he received enormous financial support from the French government; his career was also marked by ethically doubtful experiments and concealment of inconvenient details of his research.77 Nonetheless, his work, I believe, illuminates an important aspect involved in the tribulations of the inter-scientific understanding. The fact that he, a scientist from the outside of the German paradigms, could understand and evaluate the good parts of each theory (be it Linaean or Unitarian/Transformist) is a good reason to believe that a proper communication could have been established between the proponents of Kuhn and the proponents of Nageli. That not just the sociological background but also the ingrained theoretical commitments played a role in the communication breakdown of that period might well be a lesson from the history of science that could be used by present-day medicine.


I have argued that in a moderate sense, the Koch and Nageli groups worked within distinct paradigms. However, I have rejected the stronger Kuhnian claim that the terms used in different paradigms were mutually unintelligible. Focusing on the semantic aspects, I have shown that no account of incommensurability of reference can be given in this case, although, for reasons related to the sociological context and the theoretical commitments of the bacteriological schools, the two parties employed, they did not understand each other.


Kuhn (2002, pp. 13–32) et passim.


Kuhn (2002, pp. 33–37).


See Kuhn (1961, p. 65) and also Fuller (2001, p. 90).


Fleck (1981, p. 89).


See Kuhn (2002), and also Hoyningen-Huene (1993, pp. 208–212) and Fuller (2001, pp. 3–38). Kuhn’s early, more cognitivist work also considered other types of incommensurability, for instance incommensurability of perception and of values. Most contemporary commentators consider semantic incommensurability as the most challenging claim advanced by Kuhn (see e.g. Devitt 2001; Bird 2000, 2004), and due to the limited scope of the present paper, I will focus precisely on semantic incommensurability. For an insightful and interesting rendition (and development) of Kuhn’s early ideas and cognitivist slant, the reader should consult Andersen et al. (2006), especially their case study of the Copernican revolution (pp. 130–164) and their exposition of conceptual inter-relation within scientific theoretical frameworks.


See Brock (1999, pp. 72–75). For a wonderful exposition of how the idea of unbroken continuity in nature developed in the history of ideas, starting from Plato’s Timaeus, see Lovejoy (1936).


Nageli (1884) apud Mazumdar (1995, p. 39).


Strick (2002, pp. 123–125).


For instance, Albert Wernich, a Berlin physician, reported that organisms of various dysenteries could arise from normal putrefactive gut organisms, Cf. Mazumdar (1995).


For example, Hans Buchner, one of Nageli' students, reported the experimental transformation of a hay-Bacillus (a non-pathogenic kind) into Bacillus anthracis. Cf. Mazumdar (1995).


Nageli (1877) apud Mazumdar (1995, pp. 77–78).


Brock (1999, p. 183). Founded by Max von Pettenkoffer in (1878), it was the first ever established institute of hygiene in the world.


As Max von Gruber, a famous former student of Nageli and von Pettenkoffer, was arguing in 1894 ‘There is more to public health than a few laws for sewerage disposal and disinfection, important though they may be. We must remember the infected organism and its power of resistance, as well as the infecting one’ apud Mazumdar (1995, p. 94).


Cohn (1875) apud Mazumdar (1995, pp. 57–58).


Brock (1999, p. 32).


Brock (1999, pp. 94–98).


Carter (1985, pp. 355–356).


Silverstein (1996), Gradmann (2006, p. 295). The first formulation of this idea occurs in his (1878) paper on wound infections.


Koch (1880, p. 65).


Koch (1880), p.66.


Koch (1880, p. 67).


Munch (2003), Carter (1985, pp. 357–358).


One source of the “postulates” was Jacob Henle (Brock, 1999, p. 335) and, as Carter has shown, Koch also owed much to Edwin Klebs’ work (Carter, 1985, pp. 363–373).


Gradmann (2005, pp. 130–131). As Gradmann observes, the quasi-military construal of the attack of pathogenic bacteria over powerless bodies which are accordingly colonised had much to do with the external (and internal) politics of Germany at that time (Gradmann 2005, p. 343); at the same time, political rhetoric came to incorporate bacteriological and public health visions, e.g. regarding the extirpation of “invisible enemies”; (Gradmann 2000). Mendelsohn (2005) underlines the same exchange of metaphors and frameworks between politics and bacteriology (p. 91).


Koch (1890).


Gradmann (2006, p. 298). Koch misleadingly announced tuberculin as a result of his previous research on “internal disinfectants” and he had commercial interests related to the production of tuberculin. At the same time, Koch honestly believed that he had discovered the cure for tuberculosis (Gradmann 2006, pp. 297–300).


Gradmann (2006, p. 298).


Gradmann (2006, p. 299).


Fleck (1981, p. 27).


Brock (1999, pp. 293–294).


Mazumdar (1995).


Gradmann (2004, p. 476).


Ibid., p. 474.


Gradmann (2006, p. 300).


Oppenheimer and Susser (2007, p. 1240).


Mazumdar (1995, p. 82).


Brock (1999, p. 332).


See Oppenheimer and Susser (2007) and Waller (2004, pp. 62–65). Waller offers a very sympathetic view to von Pettekoffer’s position, reminding that at the time of Koch’s discovery of Vibrio Cholerae in 1883, Koch had not yet succeeded to reproduce cholera by animal inoculation (successful inoculations were realised by Waldemar Haffkine in 1885) and that between 1883 and 1890 von Pettenkoffer had gathered epidemiological data concerning moist/dry environments that was consistent with his theory. Waller does not discuss the Hamburg cholera episode though.


See for instance Hoyningen-Huene (1993, pp. 3–28), where Kuhn is classified as a Kantian anti-realist.


Devitt (2001).


See Millikan, (2000, p. 2) and Soames (2005, pp. 401–402).


See Putnam, (1973, p. 700).


However, an intensional/description variance may well be consistent with a preserved reference—for instance 'creature with a heart' and ' creature with a kidney' are two different intensions with the same extension Bird (2004, p. 65).


Bird (2000, p. 167).


Bird (2002, p. 4).


Newton-Smith (1981).


And these were not “lower” level notions, the incongruence of which could have created just “local” incommensurability, but “high” level notions, situated at the core (or apex) of the conceptual frameworks taken up by the two schools, which, if incongruous, could have engendered, in a Kuhnian perspective, broad problems of incommensurability and failure to communicate; see Andersen et al. (2006, pp. 166–167), where the claims of Kuhnian incommensurability are rightly connected with the positioning of key notions within different conceptual schemes pertaining to different scientific communities.


Fleck (1981, pp. 1–8).


See Thagard (1994).


In 1894, Max von Gruber, one of the last of Nageli's proponents, contended that there was no difference among the various types of vibrio that Linaeans had identified. Cf Mazumdar (1995, p. 90).


Bird (2004, p. 68).


Putnam's thought experiment involves a planet, Twin Earth, similar in every detail with the Earth, except that on the former planet, the liquid from rivers, lakes, etc. has a different microstructure. Cf. Putnam (1973, pp. 700–703). I’ll come back to this thought experiment aspect shortly.


Importantly, Putnam’s slogan was not supposed to rule out that descriptions can have a contribution to the fixing of the reference; see Devitt (1995, p. 60).


Certain difficulties have been immediately spotted with respect to Putnam’s causal mechanism of reference fixing. One problem was that the indexical gesture, as such, could not have been sufficiently discriminative with regard to different natural kinds, hierarchical or not, that could be instantiated by a sample (the qua-problem), or with regard to a kind sample that contained various impurities (the impurities problem); another was that the indexical gesture could have simply pointed to a natural kind that did not exist, or to something that did not represent a natural kind. The latter was arguably solved by invoking the notion of re-dubbing. The former problem was solved by way of appealing to a description that could help the causal fixing of the reference; see Brown (1998), Mellor (1977). I will shortly discuss how descriptions could be used within a causal account.


Putnam (1973, p. 702).


For instance, the use of ‘oxygen’ by the chemists of the eighteenth century would be causally connected to combustion, on this account; see Newton-Smith (1981).


See Salle (2007, pp. 611–613) Munch (2003).


In Putnam's case, O1 and O2 share the beliefs that water is important for life, rain and snow are constituted from water, sodium chloride dissolves in water etc., and are able to recognise superficial aspects of water like smell, taste, boiling and freezing points. On E the composition of water is H2O, whereas on TE is XYZ.


Gradmann (2005), Oppenheimer and Susser (2007, pp. 1239–1240, Brock (1999, pp. xii–xiii).


See Latour and Woolgar (1986) for instance.


See Devitt (1995, p. 60) who writes ‘It is important to notice some things that are not consequences [of Putnam’s slogan]. First, it is not a consequence that no aspect of meaning is in the head.[…] Second, the slogan does not commit one to a theory of direct reference […]The slogan is quite compatible with the idea that a term has a sense which determines the term's reference. What the slogan denies is that any such sense can be explained solely in terms of what is in the head. Third, it is not a consequence that any particular theory of reference is false. One could even adopt the slogan whilst holding to a description theory of natural-kind terms’ (italics added).


Kuhn (1961).


Mazumdar (1995, pp. 72–73).


Of course, it would be incorrect to say that Pasteur rejected specifically the claim of the German Unitarians since, as pointed out above, it’s not even sure whether, before 1882, he had even hear of them; Pasteur was rather responding to the related “transformist” claims that were also widespread in France. I thank one of the reviewers for making this point clear to me.


Mendelsohn (2005, p. 86).


Cohn (1854) cited by Mazumdar (1995, p. 75).


Mendelsohn (2005, pp. 85–86).


Mazumdar (1995).


Brock (1999, p. 30).


Debre (2000, pp. 306–307).


The method was employed in order to rule out that any blood corpuscles or any other filterable germ could cause the symptoms. Debre (2000, p. 307).


Bird (2004).


Bird (2004, p. 81).


Brock (1999).


Robbins (2001, pp. 76–80).


Mendelsohn (2005, p. 87).


See Geison (1995) and also Robbins (2001). Whether or not, in the case of the chicken cholera and anthrax vaccines, Pasteur used the ideas of Émile Roux or those of Jean-Joseph Toussaint, as it has been claimed (see for instance Robbins 2001, pp. 76–80) it is of no importance to my argument, which is concerned with Pasteur’s thinking, not his ethics. Even if he drew on others’ ideas, Pasteur accepted, experimented and adhered to them.



Acknowledgements suppressed for blind review.

Copyright information

© Springer Science+Business Media B.V. 2011