Concepts as Tools of Discovery1 Corinne L. Bloch In this paper, I argue that concepts are instrumental in the generation of empirical knowledge, and not merely in the integration of existing knowledge or the formation of theories. First, I discuss findings from cognitive science, showing that the act of categorization facilitates the detection of similarities and differences in a given set of data. I then proceed to analyze the case study of the concept ‘virus’, and demonstrate the ways in which the placement of these entities in a unique cognitive category, separate from the category of ‘bacterium’ (and ‘protozoa’), enabled the discovery of previously unrecognized regularities in existing data. I end by explicating the sense in which scientific concepts serve as tools of discovery.

1. Introduction In recent decades, philosophers of science have incorporated insight from cognitive science into their analysis of the dynamics of science (e.g. Andersen, Barker, and Chen 2006; Giere 1990; Nersessian 2008; Thagard 2012). These works are based on the assumption that the cognitive methods employed by scientists – despite being more sophisticated, accurate and selfreflective than ‘ordinary’ cognitive practices – are developments of such practices.2 Philosophers who have focused on scientific concepts usually studied the formation of concepts and their roles in inductive inference and in the construction of new theories, models and explanations (e.g., Holland et al. 1989; Nersessian 2008; Thagard 1992). Thus, the emphasis has been on the ways in which concepts are involved in theoretical developments – not directly empirical ones.3 Of course, the influence of the scientists’ conceptual scheme on the collection and interpretation of data has been famously argued for by Kuhn and Feyerabend (Feyerabend 1962; Kuhn 1970). This involvement of concepts in these processes, however, is often seen as mediated by the theoretical baggage that concepts carry (that is, the theoretical commitments                                                                                                                         1

This is a draft for POBAM 2014. Please do not cite or circulate.

2

For example, see Nersessian’s discussion of ‘the continuum hypothesis’ (Nersessian 2008). While I generally accept the hypothesis, one must be careful not to obliterate important differences between scientific and ordinary practices. I will attend to some of these differences later on. 3

Of course, empirical developments are often the result of theoretical ones, thus some of these works demonstrate the way in which concepts are used to advance empirical research indirectly.

 

1

involved in the formation and the use of the concept). This position is in accordance with the representational view of concepts, according to which concepts are mental representations (Fodor 2003; Margolis and Laurence 2007; Pinker 1994; Slaney and Racine 2011). Correspondingly, philosophers of science have traditionally treated concepts as vessels, or receptacles of representations (see discussion in Kindi 2012). As recently pointed out by MacLeod, this approach often disregards other possible functions of concepts beyond carrying representations – functions that are closely related to the manner in which concepts are used in investigative practices (MacLeod 2012). Indeed, several philosophers of science have moved away from the representational approach, focusing on the other roles of concepts (for example, see Arabatzis 2012; Feest 2010, 2012; MacLeod 2012). The present paper is motivated by a similar approach. It is not my intention here to argue against the representational view of concepts, or to offer a theory of concepts of my own. Rather, I intend to focus on one well-established function of concepts – the categorization of the world – and discuss its role in investigative practice. Interestingly, while categorization has been one of the most studied aspects in the psychology of concepts (reviewed in Laurence and Margolis 1999; Machery 2009; Murphy 2002), discussions that have focused on the roles of concepts in investigative practice have not examined the work that categorization itself does in the production of new empirical knowledge. To address this issue, I analyze the case study of the concept ‘virus’. I argue that concepts – in their role as mental categories – are instrumental in the generation of empirical knowledge. I begin with a discussion of findings from cognitive science, which provide the framework for my analysis of the case study. These findings point toward the way in which the act of forming mental categories may facilitate the discovery of regularities in empirical data.

2. How Can Concept-Formation Reveal Regularities? When forming concepts, we create a mental classification of entities in the world. By showing how mental categories can affect both the selection of the (perceptual and conceptual) features we attend to, as well the range of values, along those features, that we take as “the same”, I will establish that the formation of mental categories can facilitate (or obstruct) the detection of similarities and differences in empirical data.  

2

Since similarity is a relative measure, it not surprising that similarity ratings of pairs of objects are affected by the introduction of a third object. For instance, participants rated the similarity between ‘black’ and ‘white’ as low when the pair was presented in isolation, but the introduction of a third item – ‘red’ – increased the similarity rating for the pair (Medin, Goldstone, and Gentner 1993). In such cases, the introduction of a third item serves as a foil that affects the range of accepted values, along the axis of a given property (e.g., color), within which two items are accepted as similar. However, contrast items not only affect the range of values, along the axis of a specific property, within which items are considered similar (e.g., the judgment of whether the color of object A is similar to the color of object B) – they also affect the selection of properties that are taken into account in similarity judgments (e.g., the judgment of whether color should be taken into account when determining whether A and B are similar). Both philosophers and psychologists have pointed out that any two things can be taken as similar, depending on what properties one focuses on (Goodman 1972; Medin 1989; Medin, Goldstone, and Gentner 1993). As psychologist Douglas Medin put it, “a zebra and a barberpole would be more similar than a zebra and a horse if the feature ‘striped’ had sufficient weight” (Medin 1989:473). The question becomes, then, what determines the properties that are taken into account in similarity judgments, and their weights? The weighing of relevant properties for similarity judgments depends on various factors, such as expertise, goal, age, and general context (reviewed in Goldstone 1994b). One of these factors is the set of stimuli presented to subjects. Studies have shown that, when the relevant respect of similarity is not specified, it is often inferred from the general context, that is, from the set of objects under consideration (Goldstone, Medin, and Halberstadt 1997; Tversky 1977). Similarly, participants that were asked to rate overall similarity of pairs of targets, changed the weight they gave differences in size and shape according to the overall variation of these factors over the entire stimulus set used (Torgerson 1965). To account for the effects of stimulus context, Tversky and Gati have suggested the diagnosticity of features as a context-dependent component in similarity judgments models (Tversky 1977; Tversky and Gati 1978). According to these authors, a feature has a high diagnostic value in a particular context if it serves as a basis for classification in that context. That is, it has a high diagnostic value if it is not shared by all objects under study, thus can serve to differentiate one group from the rest of the set. For example, the feature “European” has no diagnostic value when all the countries presented to  

3

subjects are European, but it acquires diagnostic value when American countries are added into the set. In the latter case, the degree of similarity between the European countries in the set rises. It is possible that some of the effects of the stimulus set on overall similarity ratings are the result of a change in the weights that participants assign to detected common / differentiating attributes, for the purpose of similarity judgments (e.g., subjects may notice that two objects are similar in color, but because of the context of the task, decide not to take color into consideration for similarity judgment).4 Data indicates, however, that at least some of the effects of categories (and contrast categories) on evaluations of similarity are mediated through a change in the ease with which participants detect various similar and different properties among targets. There is a vast body of research on the effects of categorization on the perception of similarities and differences among members of a category, and among members of different categories (for review, see Collins and Olson 2014; Harnad 1987). I will only mention a small sample of these works here. Data shows that participants abstract over differences between targets that belong to the same category, while highlighting differences between targets that belong to different categories. When subjects were shown objects that were randomly labeled as belonging to one of two madeup categories (“Art from a Venusian Colony” and “Art from a Martian Colony”), features that discriminated between the two categories were selectively highlighted (Goldstone 1996). The effects of within-category assimilation and between-category accentuation has been observed when subjects were asked to evaluate various properties such as the length of straight lines (Tajfel and Wilkes 1963), colors of symbols (Goldstone 1995), daily temperatures (Krueger and Clement 1994) and grating patterns (Notman, Sowden, and Özgen 2005). Categorization yielded selective sensitization to dimensions that determine category membership, as well as local sensitization to differences in dimension values that crossed category boundaries. Additionally, categorization training has led to a decrease in ability to spot differences along a dimension that is irrelevant for categorization (Aha and Goldstone 1992; Goldstone 1994a; Hockema, Blair and Goldstone 2005). Multiple mechanisms may mediate these effects of categorization. Data from various works suggests that higher-order cognitive processes can affect the early stages of visual perception                                                                                                                         4

 

A related point is discussed in (Goldstone, Lippa, and Shiffrin 2001). 4

(reviewed in Collins and Olson 2014). Some authors, however, have argued that the early stages of vision are cognitively impenetrable, and that cognition can only affect pre-perceptual attention-allocation, or post-perceptual processes of evaluation, selection and inference (Pylyshyn 1999). Electrophysiological studies, which have a much higher temporal resolution than behavioral studies, have disagreed on whether categorical perception involves early perception or post-perceptual processes (He et al. 2014; Lu et al. 2014; Mo et al. 2011; Clifford et al. 2010; Holmes et al. 2009). The question of whether mental categories affect perception itself or processes that are distinct from (even if very closely tied to) perception, ties into the bigger debate about cognitive penetrability of perception, and has implications for traditional epistemological questions. For the purpose of this paper, however, the specific mechanism(s) by which cognition asserts its effects on the ease with which we detect similarities and differences is not important. What will be of relevance is that categorization does affect our ability to detect regularities, and furthermore – that such effects are not observed only at the perceptual level. Medin and Barsalou pointed out that some effects that have been associated with categorical perception in sensory perception categories were also observed in research on generic knowledge categories, which are usually defined by more abstract attributes (Medin and Barsalou 1990). Eiser et al. showed that effects of categories on the detection of similarities and differences were also observed on the level of non-perceptual features. Participants were asked to evaluate different statements for their degree of permissiveness toward drug use, and their evaluations tended to minimize intra-category differences and maximize inter-category differences (Eiser 1971). Furthermore, Krueger et al. observed accentuation effects while using numerical stimuli, indicating that the observed effects were not caused by modification of perception (Krueger, Rothbart, and Sriram 1989). The above data shows that the act of categorization affects the detection of similarities and differences (at various levels of abstraction) between objects. In this paper, I propose that these findings have importance for the role of concepts in investigative practice. Since the formation of a new concept involves the formation of a new mental category, it may direct one’s attention toward (or facilitate perception of) specific aspects of the phenomenon under investigation, thus enabling scientists to detect regularities that have not been previously detected. This suggests a pivotal role for concepts in scientific discovery, and another way in which scientific concepts can be viewed as tools of empirical science. In the next section, I will discuss some developments in  

5

the history of virology. I will later outline the way in which these developments can be analyzed in light of the above discussion.

3. A Case Study: The Concept of ‘Virus’ In this section, I do not attempt to provide a historical overview of the development of the virus concept.5 Rather, my goal is to discuss the possible effects of concept-formation on the observation of regularities in the data. Before I can attend to this question, however, I must determine when the concept was formed. This brings up a methodological question: what should be my criteria for establishing the formation of the virus concept? After all, as emphasized by various authors (e.g., Bos 1981; van Helvoort 1994), scientists have debated about the nature of viruses for several decades following the introduction of the concept, and it was not until much later in the 20th century that the theories about their nature came close to what we know today. My methodology here is shaped by the aspect of conceptualization that I wish to examine: the formation of mental categories. If concepts are treated as vessels of representations, it is difficult to determine the point in which a concept was accepted by the community. Do scientists holding vastly different ideas about the nature of the virus have the same concept, or do they have different concepts? In the aftermath of Kuhn and Feyerabend, several solutions have been offered, such as the causal theory developed by Kripke and Putnam (Kripke 1980; Putnam 1975). My aim here is not to offer an alternative solution to this issue. However, my focus on concepts primarily as mental categories instead of on concepts as receptacles of representation bypasses this difficulty. For my purposes here, I only need to determine when a separate category was established, without being dependent on the establishment of any particular theoretical commitments about the nature of the members of the category. My task in Section 3.1 is, therefore, to limited to discussing the point in which viruses were accepted as causal agents that should be classified under a new mental category, rather than under bacteria or other agents. In other words, I will focus on the period in which scientists accepted viruses as a kind different in nature – leaving open which of the various theories about the nature of viruses they were committed to. In Section 3.2, I will examine the                                                                                                                         5

 

For a review, see (Creager 2002; Grafe 1991; Hughes 1977). 6

effect that this acceptance might have had on the observation of regularities among instances of viral infections. Of course, even if I limit myself to the formation or acceptance of a new category as the criterion for conceptualization, without imposing further conditions about the theoretical commitments involved, I cannot treat the acceptance of the concept ‘virus’ as a single point in the history of science. As I will highlight, following the initial formation of the modern concept, it became accepted by the scientific community only gradually and with much resistance. Since my goal in this paper is to show how categories facilitated scientists’ detection of regularities, it is important that I layout the classificatory framework within which these scientists operated. For this purpose, I will briefly describe the period during which the concept became accepted by the community.6 This will provide the context for the discussion, in Section 3.2, of the reports made by several scientists when observing instances of viral infections. Before plunging into the case study, a clarification about the history of the term ‘virus’ is required. Until the first decades of the twentieth century, the term was used broadly, to designate all infectious agents, including bacteria. By the 1940’s, however, it was used more narrowly, to designate more-or-less the same group of infectious agents we now refer to as viruses. As a starting point, they can be described as submicroscopic, obligate intracellular parasites which, unlike bacteria, are unable to generate metabolic energy or to synthesize proteins. Throughout paper, unless otherwise specified, I use the term in this modern, narrow manner. 3.1. The Formation and Acceptance of the Virus Concept The germ theory, which was developed in the 19th century, maintained that infectious diseases were caused by microorganisms. The theory became widely accepted only after 1877, when independent studies by Koch and Pasteur demonstrated the etiology of anthrax, establishing a bacterial cause for the disease. A few years later, Koch formulated three conditions which were necessary to establish a bacterial etiology (Koch 1884): (i) The specific microbe must be demonstrated in all cases of the disease; (ii) The microbe must be isolated and cultured in a pure state on an artificial medium; (iii) The pure culture must produce the disease when inoculated into healthy, susceptible animals.                                                                                                                         6

Most of the historical discussion in Section 3.1 is a shorter version of a description that appeared in (Bloch 2012), which uses the same case study to discuss the roles of early scientific definitions.

 

7

With the development of techniques for the isolation of bacteria, and the formulation of criteria for the determination of bacteriological etiology, bacteriological research began to advance rapidly. The germ theory became so widespread that it was assumed that all infectious diseases were caused by bacteria. Accordingly, the agents of infectious diseases were commonly described as microorganisms detectable with light microscope, retained by bacterial filters, and cultured on artificial media. In the following years, however, scientists working on some infectious diseases failed to isolate, visualize, and grow the causal agents in vitro. For example, in the 1880’s, Louis Pasteur could not detect microscopically the infectious agent of rabies, nor could he grow it in cultures. These failings were thought to be a consequence of technical difficulties due to the small size of the infectious agents, rather than these agents being fundamentally different from the known bacteria (for review, see Hughes 1977). Pasteur wrote: “The anthrax of cattle, malignant pustule of man, is produced by a microbe; croup is produced by a microbe . . .. The microbe of rabies has not been isolated as yet, but judging by analogy, we must believe in its existence. To resume: every virus is a microbe. Although these beings are of infinite smallness, the conditions of their life and propagation are subject to the same general laws which regulate the birth and multiplication of the higher animal and vegetable beings” (cf. Pirie 1948: 329). In 1892, Dimitri Ivanovsky reported that the sap from plants infected with tobacco mosaic disease (TMD) remained infectious even after filtration through two layers of filter paper. He attributed the inability of the agent to develop on artificial media to a technical difficulty that could, in principle, be resolved. He concluded that either the infectious agent was a bacterial toxin which passed through the filter, or it was very small bacteria that passed through the pores (Ivanovsky 1892: 29-30). While the discovery of the virus is often attributed to Ivanovsky, he did not, at that point, conceive of the causal agent of the TMD as a new form of infectious agent. In 1898, Loeffler and Frosch, bacteriologists trained by Koch, published their report on footand-mouth disease. They found that the disease is transferred in lymph from epidermal vesicles. To their surprise, filtration of the lymph did not diminish its infectivity, indicating that the causal agent is able to pass through the filter. Like Ivanovsky, the authors could only conceive of two explanations for these results: either the disease is caused by a toxin that was present in the lymph or by minute bacteria that are too small to be retained by the filter. Calculating the  

8

dilution of the causal agent in their experiments, however, they contended that a soluble toxin must have a level of activity that is “truly unbelievable” in order to maintain the level of infections they observed. Thus, they argued that the infectious agent must be capable of reproduction and could not be a toxin (Loeffler and Frosch 1898). Although Loeffler and Frosch did not observe reduced infectivity of the lymph after filtration, could not grow the agent in culture and could not detect it under the microscope, they maintained their hypothesis that the agent is a small bacterium. In the same year that Loeffler and Frosch published their report, Martinus Beijerinck found, independently of Ivanovsky, that sap from plants infected with TMD remained infectious after filtration (Beijerinck 1899, 1898). Based on his finding that an indefinite number of healthy plants can be infected by the sap in series, Beijerinck concluded – as did Loeffler and Frosch with regards to the agent of foot-and-mouth disease – that the infectious agent had the capacity to multiply within living plants and therefore cannot be a toxin. Beijerinck gave this finding a different interpretation from that of his colleagues. Along with the filterability of the agent, several other results, such as the ability of the agent to diffuse through agar, indicated to him that the causal agent couldn’t be corpuscular bacteria, in the usual sense of the word. He concluded that the agent is liquid and not corpuscular. Since the virus didn’t grow in culture even when supplied with sap from healthy tobacco plants, Beijerinck further concluded that the agent is an obligatory parasite, which reproduces passively, by incorporation into the living protoplasm of the host cell. He named the agent contagium vivum fluidum (translated as ‘living germ that is soluble’), characterizing it as a non-cellular, submicroscopic, infectious agent, which can only reproduce within living cells (Beijerinck 1898; see also van Helvoort 1994: 191). Beijerinck had clearly formed a new concept, a mental category for what he viewed as a kind different in nature from bacteria (see Bos 2000: 83; Hughes 1977:57; Lwoff 1957: 241). His concept of a non-corpuscular, transmittable living agent, however, conflicted with two prevailing theories – the cell theory and the germ theory. The concept, therefore, was not immediately accepted by the scientific community (see Hughes 1977: 57-58; Waterson and Wilkinson 1978). For example, virologist Lute Bos wrote: Beijerinck's entirely new concept, launched in 1898, of a filterable contagium vivum fluidum which multiplied in close association with the host's metabolism and was distributed in phloem

 

9

vessels together with plant nutrients, did not match the then prevailing bacteriological germ theory. At the time, tools and concepts to handle such a new kind of agent (the viruses) were nonexistent. Beijerinck's novel idea, therefore, did not revolutionize biological science or immediately alter human understanding of contagious diseases (Bos 1999:675).

In a section that demonstrates how difficult it was for researchers to accept the idea of a new kind of causal agent, Thomas Rivers wrote in 1937: After a real struggle that occurred not so many years ago, certain maladies were shown to be induced either by small animals or minute plants, e.g., protozoa, fungi, bacteria and spirochetes. Indeed, the victory was so great that most workers in time began to consider that all infectious diseases, including those whose incitants had not been discovered, must be caused by agents similar to those already recognized. According to them, there could be no infections that were not caused by protozoa, fungi, bacteria or spirochetes, and to intimate that some infectious agents might be inanimate constituted heresy of the first order (Rivers 1937:1).

In the face of the failure to visualize some viruses and grow them in vitro, several works continued to provide support for the accepted view, that “filterable viruses” (i.e., infectious agents that can pass through a filter) are microbes.7 One such report was made by Edmond Nocard and Emil Roux, who researched bovine pleuropneumonia, now known to be caused by a mycoplasma and not a virus (Nocard et al. 1896). Nocard and Roux observed that, while the causal agent of bovine pleuropneumonia passes through a filter, it could be demonstrated in the microscope and – utilizing new techniques – grown in artificial cultures.8 They wrote: Discovery of the agent of pleuropneumonia virulence is not only of interest because of the difficulty overcome; its significance goes beyond. It raises the hope of also being successful in the study of other viruses whose microorganism has heretofore remained unknown. What made the determination of this microbe so difficult was, on the one hand, its extreme tenuity, and, on the other, in particular, the extremely limited culture conditions in artificial medium. Now it is quite conceivable that even smaller microbes exist, which instead of being within the limits of visibility, as is the case for this one, are beyond such limits; in other words, we can accept the fact                                                                                                                         7

When using the phrase ‘filterable virus’, the term ‘virus’ is used in its earlier, broader meaning, designating all infectious agents. Accordingly, ‘filterable viruses’ are a specific subgroup of that broad class, i.e., those infectious agents who are able to pass through a filter. 8

The paper does not discuss the filterability of the virus, but see (Hughes 1977: 65) for Nocard’s determination that the causal agent can pass through a filter.

 

10

that microbes exist which are invisible to the human eye. Well, even for these microbes, study remains possible provided a favorable culture medium is found (Nocard et al. 1896:357).

Accordingly, in a highly influential 1903 review entitled “On the so-called ‘invisible microbes’”, Roux discusses not only the causal agent of bovine pleuropneumonia, but also other infectious agent, including that of Beijerinck’s TMD, suggesting that the agent of pleuropneumonia “forms a transition between the ordinary bacteria and those which the microscope is incapable of showing” (Bos 1999: 680). Indeed, since the causal agent of bovine pleuropneumonia was almost submicroscopic, it seemed reasonable to assume that there are even smaller living cellular organisms that are invisible under the microscope. Furthermore, the fact that the causal agent could only be grown in culture using very specific bacteriological methods, gave researchers the hope that – with some technical improvements – other causal agents could eventually be isolated and grown using similar methodology.9 Thus, for the first couple of decades of the 1900’s, most researchers still viewed the difference between ‘filterable viruses’ and ordinary bacteria as primarily a matter of size rather than the pathogen being a fundamentally different type of infectious agent. Finally, after a couple of decades in which bacteriologists repeatedly tried – and failed – to grow the virus in vitro, it became difficult to sustain the view that these failures were due only to temporary technical difficulties. In 1927, Rivers wrote: In general it can be said that no worker has proved that any of the etiological agents of the [viral diseases] is susceptible to cultivation in the absence of living cells. A satisfactory explanation of the difficulty experienced in cultivating the viruses on artificial media is not easily found. Their small size alone should not make them insusceptible to cultivation. Nor does it seem to be a question of delicacy or sensitiveness, because many of them are extremely resistant to chemical and physical agents. Therefore, the viruses appear to be obligate parasites in the sense that their reproduction is dependent upon living cells (Rivers 1927: 228).

During the 1920’s, although the field was still far from achieving a consensus with regard to the nature of these agents, most scientists became convinced that they were dealing with a new class of infectious agents, whose members share several physical characteristics that go beyond

                                                                                                                        9

 

See also discussion in ((Wilkinson 1974): 213). 11

their small size.10 At that time, even those theorists who still held that viruses are minute living organisms, considered them to be different from ‘ordinary’ bacteria not only in size but also in their intimate type of parasitism, and in other related characteristics of the virus.11 The viruses were now generally viewed as a separate category – even if still mysterious. In accordance with Beijerinck’s original description – and in a way that leaves as an open question their biological nature – they were defined by the following operational (and negative) definition: viruses are invisible by ordinary microscopic methods, they fail to be retained by filters impervious to wellknown bacteria, and are unable to propagate themselves in the absence of susceptible cells (Rivers 1932:423).12 This definition differentiated viruses from bacterial and fungal agents, within the broader class of infectious agents. The concept of the virus was accepted and used by most members of the scientific community, and from now on – the focus was on elucidating their more causally fundamental properties. Over the next couple of decades, rather than continuing to attempt to study viruses using bacteriological methodology, new techniques were developed for the study of viruses, and research into their nature advanced rapidly.13

3.2. Categories and Regularities – Other Characteristics of Viruses In this subsection, I discuss several characteristics of viruses which, when viewed as common to many of them, have shed light on the unique nature of viruses and their interactions with their host cells. I will focus here on scientists’ descriptions of cellular responses observed in cases of viral infections, and then briefly mention a couple of other general characteristics of viruses. I argue that, while these characteristics had been observed in various specific viral diseases during the 19th century and the first two decades of the 20th century, they were only

                                                                                                                        10

E.g., (Rivers 1932; Stanley 1937):59. It should be noted that some researchers suggested that the category may not be a homogenous one (see (Rivers 1932): 440). 11

There were, of course, exceptions. For example, Ledingham wrote that “viruses are living organisms distinguished from the visible bacteria only by their small size” ((Ledingham 1932): 953). 12

Rivers’ description is taken to be a representative definition, reflecting the consensus that was formed in the 1920’s (see (van Helvoort 1994):186; (Norrby 2008):1118). 13

 

For review, see (Creager 2002; Hughes 1977). 12

recognized as similarities – and their significance came to light – after the formation of the ‘virus’ category.14 Inclusion bodies: Many viruses induce in their host cell the formation of inclusion bodies (IBs), which are aggregates that consist of virus particles, virus induced material, cell components, or mixtures of these elements. Observations of inclusion bodies, in cases that are now known to be viral diseases, were reported as early as 1841 (Grafe 1991). However, before viruses were classified as a separate category, these observations were not seen as instantiations of a more general phenomenon distinct from manifestation of diseases caused by other infectious agents. In the first two decades of the 20th century, many scientists reporting observations of IBs have not yet accepted the modern concept of virus. Thus, the system of categorization they had accepted included only bacteria, protozoa, fungi, etc. Equipped with these mental categories, they observed the morphology of IBs, emphasizing in their reports the similarities between IBs and parasites such as protozoa and fungi (typical descriptions of protozoa and fungi, using similar staining techniques, can be found, e.g., in Wright 1903; Wright and Craighead 1922). These similarities led many researchers to believe that what are now known to be viral IBs are instances of protozoa, rather than a phenomenon tied to a distinct kind of infectious agent (e.g., Guarnieri 1892, Bollinger 1873, cited in Rivers 1928; Calkins 1910 cited in Acton and Harvey 1911; Jackson 1920; Williams and Lowden 1906). Prowazec even labeled these IBs with the term chlamydozoa, a term that has been picked up by other workers (cited in Cowdry 1922).15 Similarities to protozoa were also found on the functional level: some protozoa are – like viruses – obligate parasites. Furthermore, these protozoa multiply only in the nucleus of the host cell, only the cytoplasm, or in both, which paralleled the various locations in the cell in which IBs have been found (Findlay and Ludford 1926).

                                                                                                                        14

Of course, today many more characteristics of viruses are known, which reveal a lot about the nature of viruses. But I will limit myself to discussion of those distinguishing characteristics that had already appeared in the data before the conceptualization of ‘virus’, but were not recognized as such until the concept was formed. 15

Other theories about the nature of IBs were suggested. For example, some argued that IBs are the product of cellular degeneration (e.g., (Acton and Harvey 1911; Goodpasture 1925)), while others have hypothesized that the bodies are composed of the colonies of parasites engulfed by products of the infected cell (e.g., von Prowazek (1907) cited in (Rivers 1928); Hartmann (1910) cited in (Acton and Harvey 1911). For my purpose here, it is important that these theories did not treat IBs as a unique new phenomenon, but rather focused on the similarities of the observed IBs to previously-classified phenomena.

 

13

Interestingly, in many individual cases, researchers were aware that the specific agent causing the appearance of observed IBs has characteristics that are unusual for protozoa. Nevertheless, they often did not consider the possibility that the IBs they have observed are caused by a new kind of agent. For example, Negri, who interpreted the bodies he observed in rabies to be protozoa, was aware that the causal agent of rabies retained its infectivity following filtration, but he interpreted it as an indication that in the agent’s life cycle, some form is small enough to pass through a filter (cited in Goodpasture 1925). Williams and Lowden were similarly well aware of the filterability of the agent. However, reflecting on the nature of the agent, they wrote: “[I]t is certain that the great majority of the bodies stand out so clearly as organisms with such definite, constant, characteristic structure and staining reactions and show so many forms similar to division forms of known Protozoa, that the picture is difficult to explain in any other way than as that of a developing organism belonging to the group Protozoa. It seems unnecessary further to consider the possibility of their being changed red blood cells or any other form of degeneration of the host tissue; and this alone is evidence in favor of their being organisms.” (Williams and Lowden 1906).

As researchers came to realize that ‘filterable viruses’ formed a new kind of infectious agents, however, they started to emphasize correlations between some of the characteristics that distinguish viral IBs from both protozoa and other cellular inclusions. Farber and Wolbach, for example, wrote in 1932, that “a mass of data has accumulated to show that a definite relation does exist between inclusion bodies and certain types of filtrable virus disease (variola, vaccima, sheep-pox, fowl-pox, molluscum contagiosum, herpes, submaxillary virus disease of guinea pigs, and so on) (Farber and Wolbach 1932). Among the typical characteristics of IBs: they are easily stained by acid dyes; a clear halo of non-stained area is visible around them; and in cases of nuclear IBs – margination of the chromatin takes place (Cowdry 1934). Additionally, IBs caused by the same virus exhibited differences in size and appearance in different species, while remaining constant for the same species. These differences among infected species were more profound than those exhibited by protozoa (the differences in size were first described in (Acton and Harvey 1911)). Other morphological characteristics that differentiated viral IBs from

 

14

protozoa were reported (Covell and Danks 1932).16 The above is a general sketch of the ways in which IBs were described before and after the formation of the virus category. Of course, this is a highly simplified description, as the acceptance of the virus as a new category did not take place simultaneously within the scientific community, thus the relation between categorization and the detection of regularities in empirical data needs to be analyzed for individual scientists, in relation to the categories that they used during this transitional period. Such relation can be demonstrated in various prominent works during the early 20th century, and I shall only provide a few cases here, in which researchers discussed their observations of the morphology of inclusion bodies caused by what is now known as cytomegalovirus (CMV): Jackson, who thought that these structures were protozoa, discussed the variation of the shape and size of these structures (Jackson 1920). She did not, however, relate these observations to similar findings, such as those reported earlier by Acton and Harvey in rabies (Acton and Harvey 1911). Her discussion overlooked the ways in which the bodies she observed are similar to other IBs now known to be caused by viruses, and at the same time – highlighted the ways in which their morphology is similar to that of protozoa. This lent itself to the interpretation of the observed differences in morphology as resulting from different developmental stages of the parasite. A different discussion was provided by Goodpasture and Talbot, who reported the pathology and histology in an infant with what is now known to be a case of CMV (Goodpasture and Talbot 1921). From the outset, they were convinced that the structure they, and others, had observed could not be protozoa, as “it would seem almost incredible that the condition could be one of protozoan infection confined to fetal and infantile life”. As they saw no evidence of an infectious agent, they concluded that the bodies they observed are products of cellular degeneration. In their paper they attempted to trace the metamorphosis of these cells, and emphasized the similarities between the pathological structures they and others have observed, and “large mononuclear wandering cells” that are found just outside the endothelial lining of small veins and capillaries in the lung. The authors did not address the morphological features that distinguish viral IBs from degenerative changes that are not caused by viruses (such differences are described, for example, in (Cowdry 1934)), and labeled the abnormal cells with                                                                                                                         16

These authors also reported that the granules observed in the IBs did not change their location or exhibit movement, providing evidence against the theory that the bodies are aggregates of individual parasites.

 

15

the name ‘cytomegalia’. Yet a different discussion of the morphology of IBs in CMV was provided by Cole and Kuttner, who, by then, took ‘filterable viruses’ to be a category separate from bacteria and protozoa (Cole and Kuttner 1926). Accordingly, they noted the various morphological similarities between the bodies they observed, and those observed in cases of other ‘filterable viruses’, such as herpes simplex, distinguishing these viruses from protozoa. Species specificity: Viruses, with some exceptions, are generally more species-specific than bacteria. This property had been reported early on in connection with reports of individual viruses. In 1898, for example, Sanarelli reported infectious myxomatosis that infects only rabbits (see Rivers 1930). In 1911, Rous published his findings that the Fowl Sarcoma (now known as Rous Sarcoma) is highly species-specific (Rous 1911). While many similar reports were available in the first two decades of the 20th century, they were not taken, as a whole, to be indicative of a new type of phenomenon. Only after the acceptance of the virus category, did scientists start treating this as a general property of ‘filterable viruses’, and connect it to other properties of these agents (e.g., Rivers 1927). Susceptibility of young, healthy cells: Beijerinck had observed, at the end of the 19th century, that infection only occurs in growing portions of the plant. Taking the virus to be a new kind of infectious agent, he saw the physiological significance of this finding, suggesting that “[w]ithout being able to grow independently, [the agent] is drawn into the growth of the dividing cells and here increased to a great degree without losing in any way its own individuality in the process” (Beijerinck 1898: 39). Since other researchers disagreed with Beijerinck’s claim that the viruses form a new kind, they did not relate similar observations in other viruses to each other. These isolated observations included early reports on various other diseases (now classified as viruses), which are similarly transmitted only by grafting (reviewed in Van der Want and Dijkstra 2006). Similarly, viral diseases usually attack insects in early stages of development, and in higher animals viruses are best exhibited in young healthy individuals. Furthermore, the infectiousness of many viruses is increased by injury – but not due to the existence of dead tissue, but rather due to the presence of young cells involved in the tissue repair. Again, many of these observations had been reported in connected with various individual viruses, before viruses were classified as a separate kind (reviewed in Rivers 1927; Grafe 1991). However, it wasn’t until the acceptance of the new category that these findings were fully related to each other, and to the other properties reported for viruses.  

16

The characteristics described above generally distinguish viruses from both bacteria and protozoa.17 They arise from the unique nature of the virus and the way in which it interacts with, and depends on, the host cell. While these characteristics had been reported for various viruses before the virus category was formed and accepted, they were not taken as a consideration in the decision to form the new category. Rather, viruses were grouped together into a category based solely on their invisibility, their filterability and their inability to grow in cell-free culture. Only after the category of ‘viruses’ has been established, were these “additional” characteristics discussed as general features of that category. My point here may seem trivial: before scientists had formed the ‘virus’ category, they did not treat characteristics such as the appearance of IBs or species specificity as characteristics of viruses. My argument, however, goes beyond this trivial point – it is that these scientists did not treat these “additional” characteristics as regularities at all. Scientists who classified viral diseases as bacteria did not discuss a subgroup of bacteria that is generally more species-specific than others. Scientists who classified IBs as protozoa did not discuss a subgroup of protozoa with morphological characteristics that are somewhat different than those observed in other protozoa. Before the category ‘virus’ was formed, scientists focused on the similarities (and did not emphasize the differences) between cases that are now classified as viral diseases, and instances of bacteria or protozoa (depending on their own classification of the phenomenon). The formation of the category ‘virus’ – based on the agent’s filterability, invisibility and inability to grow on artificial media – enabled scientists to detect additional regularities, make new connections and new distinctions, within the available data.

4. Discussion 4.1. The Role of Concepts in Detection of Regularities In light of the data from cognitive science, I suggest that the very act of grouping viruses into a single category enabled the scientists to view the “additional” characteristics of viruses as regularities, and this realization, in turn, has shed further light on the nature of viruses and                                                                                                                         17

These regularities do not provide clear-cut distinction between viruses and other causal agents. I will discuss the implications for my argument in Section 4. For now, suffice it to say that these characteristics were discussed as general characteristics of viruses after the formation and acceptance of the category (e.g. Rivers 1927, 1928, 1932). My claim here is only that it was the formation of the category that enabled this observation.

 

17

facilitated further investigation into their nature. Categorization has been shown to facilitate the detection of similarities among members of a category, as well as the detection of differences across category borders (reviewed in Collins and Olson 2014; Harnad 1987). Furthermore, various effects related to categorical perception are observed at the level of non-perceptual features (Eiser 1971) and in categories that are grouped based on abstract attributes (Medin and Barsalou 1990), supporting the suggestion that categorization may affect not only the detection of perceptual regularities (such as the morphology of IBs) but also the detection of more abstract one (such as the degree of species-specificity of the causal agent). Additionally, visual processing is influenced by verbal category labels even when the labels themselves are redundant and non-informative (Lupyan and Spivey 2010; Lupyan and Thompson-Schill 2012), and labels have been shown to facilitate categorical perception (Kikutani, Roberson, and Hanley 2008 but see Holmes and Wolff 2012). Lupyan (Lupyan 2012) suggested that language modulates aspects of conceptual representations by selectively activating perceptual features that are diagnostic of the labeled category. Therefore, it is possible that the use of explicit labels such as chlamydozoa (cited in Cowdry 1922) protozoa (e.g., Jackson 1920) or cytomegalia (Goodpasture and Talbot 1921) further strengthened the effects of categorization on the detection of regularities in the data. One may raise the following objection to my analysis of the case study: none of these “additional” properties discussed in Section 3.2 is a necessary characteristic of viruses, or one that distinguishes all viruses from bacteria and protozoa. For example, although viruses are generally more species-specific than bacteria, some bacteria show more species-specificity than some viruses. Therefore, it is possible that, while scientists had observed these regularities among instances of viruses, they simply chose not to use them as a basis for the formation of the category. To address this objection, we need to distinguish between the act of forming a new concept and the act of forming a definition.18 Thus, while it may be important that the features specified in the definition are able to clearly distinguish the items subsumed under the concept from items subsumed under different concepts, the characteristics involved in the formation of

                                                                                                                        18

As various philosophers and psychologists have argued, the classical theory of concepts, which maintained that concepts have a definitional structure, is untenable (for review see, e.g., Laurence and Margolis 1999).

 

18

the concept need not provide necessary and sufficient application conditions.19 Psychological studies have shown that people form concepts that are grounded in similarity judgments based on multiple properties of members of the category, without the requirement that any of these properties are instantiated in all and only the members of the category (Rosch 1999). For example, the property ‘has fur’ may be associated with the concept ‘cat’, serving as a basis for the initial formation of the category and as a cue for identification of instances of the category, even if people are aware of species of hairless cats. Therefore, while the general characteristics of viruses discussed in Section 3.2 may not be proper defining characteristics of viruses, as they are not exemplified in all and only viruses, they may still have a role in the formation of the concept – provided that they are recognized as regularities. And indeed –they were recognized as general characteristics of viruses following the formation of the concept (e.g., Rivers 1927, 1928, 1932). My suggestion is that the formation of the concept ‘virus’ was instrumental in this recognition. My analysis of the case study is based on the assumption that there is a continuum between the cognitive practices used in science and those used in ‘ordinary’ life (Nersessian dubbed this “the continuum hypothesis” (Nersessian 2008)). As Cheon and Machery pointed out, this assumption is either trivially true (“scientists use many ‘ordinary’ cognitive methods”) or controversial (“scientists use only ‘ordinary’ cognitive methods”) (Cheon and Machery 2010). Applying this worry to the current project, it would be prudent to consider those elements that might be unique to scientific concepts, and which might be relevant to the effects of categorization on the detection of regularities. First, one might argue that, while ‘ordinary concepts’ are based on superficial similarities, scientific concepts are based on deep causal knowledge. Psychological data has shown, however, that classification of items into ‘ordinary concepts’ utilizes causal knowledge when such knowledge is available to subjects (Ahn et al. 2000; Rehder 2003b, 2003a; Rehder and Hastie 2001; Rehder and Kim 2009).20 Thus, it seems                                                                                                                         19

I do not mean to imply that definitions are merely combinations of individually necessary and jointly sufficient conditions for applications. Indeed, I think this view of definitions prevented philosophers and psychologists from recognizing important cognitive roles of definitions (Bloch 2011, 2012). 20

Theories of concepts differ on whether categorization is based on similarity or on causal knowledge (see Machery 2009). While I think some elements of the similarity-based approach are helpful, and have relied here on some empirical work provided by this camp, I believe similarity is not sufficient to explain categorization. I am not able to discuss it at length here, but in Sections 4.2 and 4.3 I touch very briefly on the limitations of a view that is based solely on similarity.

 

19

that the difference between scientific and ordinary concepts is a difference in degree, not in kind. One interesting difference between scientific and ordinary concepts – which might have a bearing on the effects of categorization on the detection of regularities – is the fact that scientists form their own categories. Rather than learning new categories by deferring to authority, and treating those categories as constant, scientists often form their own categories, or change existing categories. Scientists are the authority. It is therefore reasonable to wonder whether scientific categories will have the same effect as ‘ordinary categories’ on the attention to (or perception of) certain regularities. Foroni and Rothbart showed that assimilation and accentuation effects of categorization were evident even when categories were generated by the participants themselves (Foroni and Rothbart 2011). Thus, even though scientists might be more open to revision of their categories in light of new data – the categories that they have accepted might still serve as tools for the detection of similarities and differences in their data. 4.2. Concepts as Tools of Discovery I should now explicate my usage of the term ‘tools’ to describe the role concepts play in empirical investigation. I borrow the notion from Uljana Feest, who argued that concepts have a productive role in empirical research, and that they thus serve as tools for the generation of knowledge (Feest 2010). Feest argued that operational definitions provide scientists with instructions about experimental interventions, and I discussed the ways in which such definitions – despite their dependence on the available experimental practices – can be used to refer to what scientists take to be a stable phenomenon, whose existence is independent of these practices (Bloch 2012). My focus here, though, is on a more basic aspect of conceptualization – the formation of mental categories. I suggested that these mental categories themselves can also be viewed as tools, as they serve the scientists as a device that highlights certain regularities in the data. In emphasizing the effects of categorization itself, over and above the theoretical commitments or expectations about the nature of the phenomena subsumed under the concept, it may seem as if I require a complete distinction between the two aspects of the formation of concepts. It is important to note, however, that even initial categorization – based on similarity with respect to superficial, rather than causally fundamental, characteristics – seems to depend on some theoretical commitments about causal structure, which determine which are the relevant  

20

characteristics that are taken into account in categorization decisions (Ahn et al. 2000; Rehder 2003b, 2003a; Rehder and Hastie 2001; Rehder and Kim 2009). To put it differently, categorization itself is a theoretical commitment. Therefore, it is impossible to distinguish the effects of mental categorization from the effects of the theoretical commitments that such categorizations reflect. For example, in examining Jackson’s report on the morphology of the bodies she had observed and their similarities to protozoa, we cannot distinguish the effects of her theoretical commitments (such as the knowledge about the life cycle of protozoa) from the effects of the categorization she has accepted. I do not suggest that we should ignore the possible effects of theoretical knowledge and expectations on the detection of similarities, and the roles they may have played in the above case study. Studies with artificial categories, however, have shown that accentuation and assimilation effects are observed even for arbitrary perceptual categories (e.g. grating patterns (Notman, Sowden, and Özgen 2005)), which are detached from any additional theoretical commitments. Thus, accounts of the role of conceptualization in the observations of similarities in empirical data should take into account these cognitive mechanisms. 4.3. Concepts as Tools of Discovery In light of the results from cognitive science and the description of the case study, one might object that concepts are means of distortion rather than tools of discovery. It seems that our categories skew our perception of (or attention to) similarities. If I argue that the concept of ‘virus’ facilitated some researchers’ observations of the distinguishing characteristics of viral infections (e.g., Acton and Harvey 1911), then I must also concede that, lacking a proper concept of ‘virus’, the concept of ‘protozoa’ had prompted others to overlook such characteristics (e.g., Jackson 1920). There are several ways to read such an objection. It could be read as conveying the commitments of scientific realism: (1) there are real, mind-independent kinds out there; (2) scientific concepts should map onto these kinds (e.g., Ellis 2001). Accordingly, when scientific concepts do not map onto these kinds – they serve as means of distortions. I have argued elsewhere that at least some kinds that scientists take as real, are not eternal, mind-independent kinds, nor do scientists assume that they are such kinds (Bloch 2011). Another reading of the above objection, which is more relevant to the focus of the current paper, conveys a commitment  

21

about the relation between categories and overall similarities. This reading implies that: (1) the degree of overall similarity between members of a category is higher than the degree of overall similarity between members of different categories, such that categories maximize in-category similarity and inter-category differences; (2) scientific concepts should map onto these categories. Therefore, when scientific concepts do not map onto these categories – they serve as means of distortions. Variations of this view have been discussed not only by philosophers of science, but also by psychologists, when applied to ‘ordinary’ concepts. For example, Rosch’s working assumption is that “in the perceived world, information-rich bundles of perceptual and functional attributes occur that form natural discontinuities, and that basic cuts in categorization are made at these discontinuities” (Rosch 1973: 192; see also Rosch 1999). When Rosch’s view is applied to scientific concepts, it implies that there is, in principle, a way to calculate overall similarities based on all the perceived characteristics of phenomena. This assumption, however, is problematic, and it has been shown that classification based on maximizing overall in-category similarity, rather than similarity in specific respects, does not result in stable categories. The early numerical taxonomists, for example, attempted to classify organisms in a theory-free, inductive method. In order to calculate similarity, they used – in equal weights – all of the organisms’ legitimate biological features (Hull 1990).21,22 One problem was that there seemed to be an indefinite number of ways to subdivide organisms into characters, and without any selection criteria – all subdivisions were equally plausible. Based on the assumption that the world is actually carved up into categories with high overall similarity, they hypothesized that as more characters are taken into account, stable categories would ensue. Contrary to these expectations, however, the categories continued to change as new data was added. The notion of overall similarity failed. As I argued earlier, this is not merely a result of a practical difficulty of discovering the overall similarity – but rather, a result of the relative nature of the very notion of similarity. There is no overall similarity – only similarity with respect to specified properties (Goodman 1972; Medin 1989; Medin, Goldstone, and Gentner 1993). Accordingly, Medin writes “[a]ttempts to describe category structure in terms of similarity will prove useful only to the extent that one specifies which principles determine what is to count as a relevant property and which principles determine the importance of particular                                                                                                                         21

Of course, “legitimate biological features” already implies some degree of judgment of relevance.

22

This example is discussed in (Griffiths 1997):178.

 

22

properties” (Medin 1989:474). Since similarity as such cannot serve as a basis for categorization, maximization of in-category similarity (and inter-category difference) cannot serve as a standard for proper categorization. Accordingly, the problem with the concept of ‘protozoa’, for example, when applied too broadly to include instances we now identify as viral IBs, cannot be that it does not correspond to a measure of overall similarity, or that it impeded the detection of regularities that contribute to such a measure. Even if one is not committed to realism about kinds or to categories that reflect overall similarity, however, one could still argue that the above case study shows that concepts can distort scientists’ observation of important regularities in the data, and that the studies about categorical perception provide us with a mechanism for such distortion. Even if there is no such thing as overall similarity but only similarities with respect to specific attributes, the detection of some similarities and differences would shed light on causal structures in the world, while the detection of others would be far less useful for this purpose. While some mental categories seem to facilitate the detection of important regularities, others seem to obstruct it. I agree. Abandoning the representational view of concepts enables us to say that concepts are neither true nor false, but that does not mean there is no room for normativity. Proper conceptualization facilitates empirical research (among other things); improper conceptualization impedes it. Taking seriously the notion of concepts as tools, we need to measure their utility against the goals of such research. Of course, this would be the case for any scientific tool. The resolution of the microscope, for example, determines which similarities and differences are observed, and which are overlooked. The same goes for using a dye to selectively stain different types of tissue. As Feest pointed out, concepts “are tools for knowledge generation, and like other tools they can be adapted or discarded in the process” (Feest 2010: 174). We can see this process clearly in the case of the virus concept. The old concept of virus, which referred to all infectious agents, was based on actual similarities among these agents, such as their ability to multiply within the host (as opposed to toxins, for example). In that context, the concept served as a proper tool for empirical study. At the turn of the 20th century, however, following the formulation of Koch’s postulates, an additional distinction was necessary for empirical investigation to continue – the distinction between ‘filterable viruses’ and bacteria. The concept needed to be modified (or subdivided) to reflect this distinction. Once the new concept of virus was formed, based on filterability, invisibility, and inability to grow in culture, the study  

23

into the nature of viruses began to advance rapidly, both because of the development of new techniques that no longer relied on Koch’s postulates, and because additional regularities came into light. This reflects the constant feedback relationship between the division of the world into categories, and the theories scientists form about phenomena classified by these categories. Scientists form concepts based on initially observed similarities, within the context of available causal theories. These concepts are then used to facilitate further empirical research, and integrate the new findings with existing knowledge. As a result of new findings, classification may be modified, and so on. I argued that the very act of grouping phenomena into a single category and distinguishing them from other phenomena, enables, among other things, the discovery of previously unrecognized regularities in empirical data. Thus, concepts themselves – apart from any experimental procedures that they facilitate – serve in the production of new knowledge. Concepts are, in this deep sense, tools of discovery.

References Acton, Hugh W., and W.F. Harvey. 1911. "The nature and specificity of Negri bodies." Parasitology no. 4 (3):255-272. Aha, D. W., and Robert L. Goldstone. 1992. Concept learning and flexible weighting. Paper read at the Fourteenth Annual Conference of the Cognitive Science Society. Ahn, W., Nancy S. Kim, Mary E. Lassaline, and Martin J. Dennis. 2000. "Causal status as a determinant of feature centrality." Cognitive Psychology no. 41:361-416. Andersen, Hanne, Peter Barker, and Xiang Chen. 2006. The Cognitive Structure of Scientific Revolutions. West Nyack, NY, USA: Cambridge University Press. Arabatzis, Theodore. 2012. "Experimentation and the meaning of scientific concepts." In Scientific Concepts and Investigative Practice, edited by Uljana Feest and Freidrich Steinle, 149-166. De Gruyter. Beijerinck, Martinus W. 1898. "Concerning a contagium vivum fluidum as cause of the spot disease of tobacco leaves. English translation published in: (1942)." In Phytopathological Classics, 33-54. St. Paul, MN: American Phytopathological Society. ———. 1899. "On a Contagium vivum fluidum causing the Spotdisease of the Tobacco-leaves." KNAW, Proceedings, Amsterdam no. 1:170-176. Bloch, Corinne L. 2011. "Scientific kinds without essences." In Properties, powers and structures: Issues in the metaphysics of realism, Routledge Studies in Metaphysics, edited by A. Bird, Ellis, B., Sankey, H., 233-255. Routledge. ———. 2012. "Early concepts in investigative practice— the case of the virus." In Scientific Concepts and Investigative Practice, edited by Uljana Feest and Freidrich Steinle, 191218. de Gruyter.  

24

Bos, Lute. 1981. "Hundred years of Koch's Postulates and the history of etiology in plant virus research." Netherlands Journal of Plant Pathology no. 87:91-110. ———. 1999. "Beijerinck's work on tobacco mosaic virus: Historical context and legacy." Philosophical transactions of the Royal Society of London. Series B, Biological sciences no. 354:675-685. Bos, Lute 2000. "100 years of virology: from vitalism via molecular biology to genetic engineering." Trends in Microbiology no. 8 (2):82-87. Cheon, Hyundeuk, and Edouard Machery. 2010. "Creating ccientific concepts, by Nancy J. Nersessian." Mind no. 119 (475):838-844. Clifford, Alexandra, Amanda Holmes, Ian Davies, and Anna Franklin. 2010. "Color categories affect pre-attentive color perception." Biological Psychology no. 85 (2):275-282. Cole, Rufus, and Ann G. Kuttner. 1926. "A filterable virus present in the submaxillary glands of guinea pigs." The Journal of experimental medicine no. 44 (6):855-873. Collins, Jessica A., and Ingrid R. Olson. 2014. "Knowledge is power: How conceptual knowledge transforms visual cognition." Psychonomic bulletin & review:1-18. Covell, W. P., and W. B. C. Danks. 1932. "Studies on the nature of the Negri body." The American Journal of Pathology no. 8 (5). Cowdry, E. V. 1922. "The supravital staining of vaccine bodies." The Journal of experimental medicine no. 36 (6):667-684. ———. 1934. "The problem of intranuclear inclusions in virus diseases." Archives of Pathology no. 18:527-542. Creager, Angela N.H. 2002. The life of a virus: Tobacco mosaic virus as an experimental model, 1930-1965: University of Chicago Press. Eiser, J. R. 1971. "Enhancement of contrast in the absolute judgment of attitude statements." Journal of Personality and Social Psychology no. 17 (1):1-10. Ellis, Brian D. 2001. Scientific essentialism, Cambridge studies in philosophy. Cambridge, U.K. ; New York: Cambridge University Press. Farber, S. and B. Wolbach 1932. "Intranuclear and cytoplasmic inclusions (“protozoan-like bodies”) in the salivary glands and other organs of infants." The American Journal of Pathology 8(2): 123135. Feest, Uljana. 2010. "Concepts as tools in the experimental generation of knowledge in cognitive neuropsychology." Spontaneous Generations: A Journal for the History and Philosophy of Science no. 4 (1):173-190. ———. 2012. "Exploratory experiments, concept formation, and theory construction in psychology." In Scientific Concepts and Investigative Practice, edited by Uljana Feest and Freidrich Steinle, 167-190. De Gruyter. Feyerabend, Paul. 1962. "Explanation, reduction, and empiricism." In Minnesota studies in the philosophy of science 3, edited by H. Feigl and G. Maxwell, 28–97. Minneapolis: University of Minnesota Press. Findlay, G. M., and R.J. Ludford. 1926. "The ultra-microscopic viruses: I.—Cell inclusions associated with certain ultra-microscopic diseases—A pictographic review." British Journal of Experimental Pathology no. 7 (5). Fodor, Jerry A. 2003. Hume Variations. Oxford: Oxford University Press. Foroni, Francesco, and Myron Rothbart. 2011. "Category boundaries and category labels: When does a category name influence the perceived similarity of category members?" Social Cognition no. 29 (5):547-576.  

25

Giere, Ronald N. 1990. Explaining Science: A Cognitive Approach. Edited by David L. Hull, Science and Its Conceptual Foundations. Chicago and London: The University of Chicago Press. Goldstone, Robert L. 1994a. "Influences of categorization on perceptual discrimination." Journal of Experimental Psychology: General no. 123 (2):178. ———. 1994b. "The role of similarity in categorization: providing a groundwork." Cognition no. 52:125–157. ———. 1995. "Effects of categorization on color-perception." Psychological Science no. 6 (5):298-304. ———. 1996. "Isolated and interrelated concepts." Memory & Cognition no. 24 (5):608-628. Goldstone, Robert L., Yvonne Lippa, and Richard M. Shiffrin. 2001. "Altering object representations through category learning." Cognition no. 1:27-43. Goldstone, Robert L., Douglas L. Medin, and Jamin Halberstadt. 1997. "Similarity in context." Memory and Cognition no. 25 (2):237-255. Goodman, Nelson. 1972. "Seven strictures on similarity." In Problems and projects, edited by Nelson Goodman, xii, 463. Indianapolis: Bobbs-Merrill. Goodpasture, Ernest W. 1925. "A Study of rabies, with reference to a neural transmission of the virus in rabbits, and the structure and significance of Negri bodies." The American journal of pathology no. 1 (6). Goodpasture, Ernest W., and F.B. Talbot. 1921. "Concerning the nature of "protozoan-like" cells in certain lesions of infancy." American Journal of Diseases of Children no. 21 (5):415425. Grafe, Alfred. 1991. A History Of Experimental Virology. New York: Springer-Verlag. Griffiths, Paul E. 1997. What Emotions Really Are: The Problem Of Psychological Categories, Science And Its Conceptual Foundations. Chicago, Ill.: University of Chicago Press. Harnad, Stevan. 1987. "Category induction and representation." Cognition and Brain Theory no. 5:535-565. He, Xun, Christoph Witzel, Lewis Forder, Alexandra Clifford, and Anna Franklin. 2014. "Color categories only affect post-perceptual processes when same-and different-category colors are equally discriminable." cortex 16, no. 17: 19." Journal of Optical Society of America A no. 31 (4):A322-A331. Hockema, Stephen A., Mark R. Blair, and Robert L. Goldstone. 2005. "Differentiation for novel dimensions." In Proceedings of the twenty-seventh annual conference of the cognitive science society, 953-958. Holland, John H., Keith J. Holyoak, Richard E. Nisbett, and Paul R. Thagard. 1989. Induction: MIT Press. Holmes, Amanda, Anna Franklin, Alexandra Clifford, and Ian Davies. 2009. "Neurophysiological evidence for categorical perception of color." Brain and Cognition no. 69 (2):426-434. Holmes, K. J., and P. Wolff. 2012. "Does categorical perception in the left hemisphere depend on language?" Journal of Experimental Psychology: General no. 141:439-443. Hughes, Sally Smith. 1977. The Virus: A History of The Concept. London, New York: Heinemann Educational Book; Science History Publications. Hull, David L. 1990. Science as a Process: An Evolutionary Account of the Social and Conceptual Development of Science, Science and Its Conceptual Foundations.

 

26

Ivanovsky, Dmitrii. 1892. "Concerning the mosaic disease of the tobacco plant." English translation printed in: (1942)." In Phytopathological Classics, 27-30. St. Paul, MN: American Phytopathological Society. Jackson, L. 1920. "An intracellular protozoan parasite of the ducts of the salivary glands of the guinea-pig." The Journal of Infectious Diseases no. 26 (4):347-351. Kikutani, M., D. Roberson, and J. R. Hanley. 2008. "What’s in the name? Categorical perception for unfamiliar faces can occur through labeling." Psychonomic Bulletin & Review no. 15:787-794. Kindi, Vasso. 2012. "Concept as vessel and concept as use." In Scientific Concepts and Investigative Practice, 23-46. Koch, Robert. 1884. "The etiology of tuberculosis." English translation of exerpts of the paper published in: (1999)." In Milestones in Microbiology: 1546 to 1940, edited by Thomas D. Brock, 116-118. Washington, DC: The American Society for Microbiology. Kripke, Saul A. 1980. Naming and Necessity. Cambridge, Mass.: Harvard University Press. Krueger, J., and R. W. Clement. 1994. "Memory based judgments about multiple categories: A revision and extension of Tajfel’s accentuation theory." Journal of Personality and Social Psychology no. 67:35-47. Krueger, Joachim, Myron Rothbart, and N. Sriram. 1989. "Category learning and change: Differences in sensitivity to information that enhances or reduces intercategory distinctions." Journal of Personality and Social Psychology no. 56 (6):866. Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. 2nd ed. Chicago: University of Chicago Press. Laurence, Stephen, and Eric Margolis. 1999. "Concepts and cognitive science." In Concepts : Core readings, edited by Eric Margolis and Stephen Laurence, 3-81. Cambridge, Mass.: MIT Press. Ledingham, J. C. G. 1932. "A Paper on tissue changes in virus diseases." The British Medical Journal no. 2 (3751):953-957. Loeffler, Friedrich J., and Paul Frosch. 1898. "Report of the commission for research on the footand-mouth disease." In Milestones in Microbiology: 1546 to 1940. Published in (1999), edited by Thomas D. Brock, 149-153. Washington, DC: American Society for Microbiology. Lu, Aitao, Ling Yang, Yanping Yu, Meichao Zhang, Yulan Shao, and Honghong Zhang. 2014. "Event-related potentials reveal linguistic suppression effect but not enhancement effect on categorical perception of color." Scandinavian Journal of Psychology. Lupyan, Gary. 2012. "Linguistically modulated perception and cognition: The label-feedback hypothesis." Frontiers in Psychology no. 3:54. Lupyan, Gary, and Michael J. Spivey. 2010. "Redundant spoken labels facilitate perception of multiple items." Attention, Perception, & Psychophysics no. 72 (8):2236-2253. Lupyan, Gary, and S. L. Thompson-Schill. 2012. "The evocative power of words: Activation of concepts by verbal and nonverbal means." Journal of Experimental Psychology: General no. 141:170-186. Lwoff, Andre 1957. "The concept of virus." Journal of General Microbiology no. 17:239-253. Machery, Edouard. 2009. Doing without Concepts: Oxford University Press. MacLeod, Miles. 2012. "Rethinking scientific concepts for research contexts: The case of the classical gene." In Scientific Concepts and Investigative Practice, edited by Uljana Feest and Freidrich Steinle, 47-74. De Gruyter.  

27

Margolis, Eric, and Stephen Laurence. 2007. "The ontology of concepts—Abstract objects or mental representations?" Nous no. 41 (4):561–593. Medin, D. L. 1989. "Concepts and conceptual structure." Am Psychol no. 44 (12):1469-81. Medin, Douglas L., and Lawrence W. Barsalou. 1990. "Categorization processes and categorical perception." In Categorical Perception: The Groundwork of Cognition, edited by Stevan R. Harnad, 455-490. Cambridge University Press. Medin, Douglas L., Robert L. Goldstone, and Dedre Gentner. 1993. "Respects for similarity." Psychological Review no. 100 (2):254-278. Mo, Lei, Guiping Xu, Paul Kay, and Li-Hai Tan. 2011. "Electrophysiological evidence for the left-lateralized effect of language on preattentive categorical perception of color." Proceedings of the National Academy of Sciences no. 108 (34):14026-14030. Murphy, Gregory L. 2002. The Big Book of Concepts. Cambridge, Mass. ; London: MIT Press. Nersessian, Nancy J. 2008. Creating Scientific Concepts. Cambridge MA: MIT Press. Nocard, E. I. E., E. Roux, Borrel, Salimbenient, and Dujardin-Beaumetz. 1896. "The microbe of pleuropneumonia." English translation published in: (1990)." Reviews of Infectious Diseases no. 12 (2):354-358. Norrby, Erling. 2008. "Nobel Prizes and the emerging virus concept." Archives of Virology no. 153: 1109–1123. Notman, Leslie A., Paul T. Sowden, and Emre Özgen. 2005. "The nature of learned categorical perception effects: A psychophysical approach." Cognition no. 95 (2):B1-B14. Pinker, Steven. 1994. The Language Instinct: The New Science of Language and Mind. London: Penguin. Pirie, NW. 1948. " Development of ideas on the nature of viruses." British Medical Bulletin no. 5(4-5):329-333. Putnam, Hilary. 1975. "The meaning of 'meaning'." In Language, mind, and knowledge, Minnesota studies in the philosophy of science ; v. 7, edited by Keith Gunderson, vii, 424. Minneapolis: University of Minnesota Press. Pylyshyn, Zenon. 1999. "Is vision continuous with cognition?: The case for cognitive impenetrability of visual perception." Behavioral and brain sciences no. 22 (3):341-365. Rehder, Bob. 2003a. "Categorization as causal reasoning." Cognitive Science no. 27:709-748. ———. 2003b. "A causal-model theory of conceptual representation and categorization." Journal of Experimental Psychology: Learning, Memory and Cognition no. 29 (6):11411159. Rehder, Bob, and Reid Hastie. 2001. "Causal knowledge and categories: The effects of causal beliefs on categorization, induction, and similarity." Journal of Experimental Psychology: General no. 130 (3):323-360. Rehder, Bob, and ShinWoo Kim. 2009. "Classification as diagnostic reasoning." Memory and Cognition no. 37 (6):715-729. Rivers, Thomas M. 1927. "Filterable viruses: A critical review." Journal of Bacteriology no. 14 (4):217-258. ———. 1928. "Some general aspects of pathological conditions caused by filterable viruses." The American journal of pathology no. 4 (2):91-124. ———. 1930. "Infectious myxomatosis of rabbits: Observations on the pathological changes induced by virus myxomatosum (Sanarelli)." The Journal of Experimental Medicine no. 51 (6):965. ———. 1932. "The Nature of Viruses." Physiological Reviews no. 12 (3):423-452.  

28

———. 1937. "Viruses and Koch's Postulates." Journal of Bacteriology no. 33 (1):1-12. Rosch, Eleanor. 1973. "On the internal structure of perceptual and semantic categories." In Cognitive development and the acquisition of language, edited by Timothy E. Moore, xii, 308. New York: Academic Press. ———. 1999. "Principles of categorization." In Concepts: Core reading, edited by Eric Margolis, and Stephen Laurence. Cambridge, Massachusetts; London, England: MIT Press. Rous, Peyton. 1911. "A sarcoma of the fowl transmissible by an agent separable from the tumor cells." The Journal of Experimental Medicine no. 13 (4):397-411. Slaney, Kathleen L., and Timothy P. Racine. 2011. "On the ambiguity of concept use in psychology: Is the concept “concept” a useful concept?" Journal of Theoretical and Philosophical Psychology no. 31 (2):73-89. Stanley, Wendell M. 1937. "Crystalline tobacco-mosaic virus protein." American Journal of Botany no. 24:59-68. Tajfel, Henri, and Alan L. Wilkes. 1963. "Classification and quantitative judgement." British Journal of Psychology no. 54 (2):101-114. Thagard, Paul. 1992. Conceptual Revolutions: Princeton University Press ———. 2012. The Cognitive Science of Science: Explanation, Discovery, and Conceptual Change: The MIT Press. Torgerson, W. S. 1965. "Multidimentional scaling of similarity." Psychometrica no. 30 (4):379393. Tversky, Amos. 1977. "Features of similarity." Psychological Reviews no. 84 (4):327-352. Tversky, Amos, and Itamar Gati. 1978. "Studies of similarity " In Cognition and categorization, edited by Eleanor Rosch and Barbara L. Lloyd. Hillsdale, NJ: Lawrence Erlbaum. Van der Want, J. P. H., and J. Dijkstra. 2006. "A history of plant virology." Archives of virology no. 151 (8):1467-1498. van Helvoort, Ton 1994. "History of virus research in the twentieth century: The problem of conceptual continuity " History of Science no. 32 (2):185-235. Waterson, A. P., and Lise Wilkinson. 1978. History of Virology. London, New York, Melbourne: Cambridge University Press. Wilkinson, Lise. 1974. "The development of the virus concept as reflected in corpora of studies on individual pathogens: 1. Beginnings at the turn of the century." Medical History no. 18 (3):211-221. Williams, Anna Wessels, and May Murray Lowden. 1906. "The etiology and diagnosis of hydrophobia." Journal of Infectious Diseases no. 3 (3):453-483. Wright, J. Homer. 1903. "Protozoa in a case of tropical ulcer (“Delhi sore”)." The Journal of Medical Research no. 10 (3):472. Wright, J. Homer, and Eugene M. Craighead. 1922. "Infectious motor paralysis in young rabbits." The Journal of Experimental Medicine no. 36 (1):135-140.

 

29

POBAM 2014 typos fixed.pdf

tools of discovery. 1. Introduction. In recent decades, philosophers of science have incorporated insight from cognitive. science into their analysis of the ...

277KB Sizes 4 Downloads 163 Views

Recommend Documents

POBAM Submission.pdf
Page 1 of 14. Page 1 of 14. Limiting Cases for the New Mechanists: Intralevel Causation is Insufficient for Modeling. Complex Biological Systems. Abstract. Mechanistic accounts of explanation pervade the philosophy of science. Typically, however,. su

POBAM paper (02-1.05.14).pdf
There was a problem previewing this document. Retrying... Download. Connect more apps... Try one of the apps below to open or edit this item. POBAM paper ...

POBAM Barker & Velasco, Fixing Evolutionary Groups
Fixing Evolutionary Groups. Matthew J. Barker ([email protected]) and Joel D. Velasco ([email protected]). Abstract. We argue for a new conventionalism about many kinds of evolutionary groups, including clades, cohesive units, and popula

Nicholson & Gawne - WPhilBioW (POBAM draft) (1).pdf
... discussed under the heading 'Pre-History of Philosophy of Biology'. (Griffiths 2008). Page 3 of 20. Nicholson & Gawne - WPhilBioW (POBAM draft) (1).pdf.

[ March 13, 2014 ] [No. 30 s. 2014] CONSTITUTING THE 2014 ...
[ March 13, 2014 ] [No. 30 s. 2014] CONSTITUT ... HOOL REPAIR-REHAB UNDER SCHOOL MOOE FUND .pdf. [ March 13, 2014 ] [No. 30 s. 2014] CONSTITUT .

JUNE 2014
THE PEORIA CATHOLIC HIGH SCHOOL PARENT NEWSLETTER .... “Geniverse is free, web-based software for high school Biology that engages students in ...

Jun 2014
The Board selects Comerica Bank as its financial institution. b. The Board ... regulations as published in the MHSAA's BULLETIN and current. HANDBOOK and shall .... of a term note (the “Note”) which will be financed in compliance with the ...

December 2014
conversion of one isomer into another. (vii) The separation of liquid ... (vii) MAP of minimally processed foods. 5. (a) Discuss the functional properties and food.

2014. 02. March 2014.pdf
office would be able to build capacities at the. grassroots level. ... EDITORIAL STAFF. Sarah Bautista □ Maria Celina Bernardo .... March 2014.pdf. 2014. 02.

2014 - AnnualReports.com
The Galileo project and the social media platform have been embraced by employees, and project ... social media platform were implemented in fiscal 2014 and will continue to be ...... Operating leases. Finance leases. Total. 2015. 57.6. 0.6. 58.2. 20

SYRCoSE 2014
software engineering. The topics of interest include but are not restricted to: • Software Development Processes. • Object-Oriented Analysis and Design.

June 2014
Term-End Examination. June, 2014. MVP -002 : PG ... (iii) Warehousing. (iv) Tariff Value. ... under this Act. (b) Discuss 3 categories of licences which were 3+2=5.

RAWSN 2014
International Workshop on RFID and Adaptive Wireless Sensor Networks: Mobility ... Radio Frequency Identification (RFID) technology enables the non-contact, ...

June 2014
BACHELOR OF COMPUTER APPLICATIONS. Cr). (PRE - REVISED) cr). Term-End Examination. June, 2014. CS- 74 : INTRODUCTION TO INTERNET.

2014 - AnnualReports.com
!ndigo Enrich Your Life, Chapters, !ndigo, Coles and indigo.ca are trade marks ... strong performing titles, this year was one in which we had no single breakout book. .... to high-needs elementary schools across the country through donations.

Jun 2014
Motion: BE IT RESOLVED that the Public School Academies of Detroit approves the agenda as presented and approves the Minutes from the April 17, 2014 ...

June 2014
1863. AFW-1(E)/BHDA-101/AFW-1(H). Total No. of Questions : 5+5]. [Total No. of Printed Nes 4. BACHELOR'S DEGREE PROGRAMME. Term-End Examination.

June 2014
some school level activities, which can be infused with this model. ... (741-.A-1-.74.fl t di-19114-14' c11. 3111.Tffi-dT (effectiveness). cRc;tiul f\R! MESE-062. 4 ...

June 2014
4. (b) Find the components of 0 under the usual metric. 2. (c) State Monotone Convergence Theorem. Verify the theorem for {fn) where fn : R —> R is defined by.

June 2014
CS-75 : INTRANET ADMINISTRATION. Time : 2 hours ... (a) What is WAP (Wireless Application 5 protocol) ? ... (c) What is Virtual Private Network ? Explain. 5.

December 2014
13-11 (Chow) IT N1111 "WE. 31-T471 fa-lk 3117 pit 41IIM #9417T. 7L- F. 4R.nr,fra af-d-rfff 71--k. -FT. 8. ralcil -14. HMO yi=f3xi+cit. 12. MW111 Gre2 t I 1:Tri rilr\IR.

SYRCoSE 2014
Software Development and Analysis Tools. • Software Engineering Education. Paper Submission. We accept papers authored by students, PhD students or ...

June 2014
annual turnover of the company is around. Rs. 150 crores and its capacity ... the accident occurred during and in the course of employment, the Management ...

June 2014
CS-612 : PC SOFTWARE APPLICATION SKILLS. Time : 2 hours. Maximum Marks : 60 ... features of any one search engine. 4. (a) What is a Chart Wizard ?