Extensive empirical research has shown that many patients think they are entitled to understand basic health concepts like disease, illness and sickness in ways that do not correspond to the understanding health workers have. Patients’ beliefs about the meaning of the concepts are heavily shaped by their idiosyncratic perspectives and social-cultural frameworks of interpretation (Macklin, 2006; Helman, 2007; Dutta, 2008; Hogg & Holland, 2010) .
This diversity of understanding constitutes a challenge for anyone who attempts to analyse the meaning of disease and illness: Analyses of these concepts have traditionally been univocal; the aim has been to develop general concept definitions (Marinker, 1975; Nordenfelt, 1987; Boorse, 1997; Worrall & Worrall, 2001) . Typically, disease has been described as “negative bodily occurrences as conceived of by the medical profession, illness as negative bodily occurrences as conceived of by the individual, and sickness as such occurrences as conceived of by the society” (Hofmann, 2002: p. 150) .
Various definitions have been developed within this threefold conceptual framework, but they all share the assumption that there is one correct analysis of each of the three concepts. So how, then, can the analyses incorporate the fact that the concepts seem to have multiple meanings in our linguistic community? Is it correct to assume that the concepts have one normative definition, grasped by a group of experts? Or is it necessary to take the diversity of understanding into account in overall analyses of each concept?
This article discusses the possibility of defining basic health concepts in the light of an influential producer-consumer distinction from philosophy of mind and language. According to this distinction, if a patient is willing to defer to a health worker’s explanation of a health term, he is a consumer of the health worker’s understanding (Burge, 1979; Goldberg & Pessin, 1996; Putnam, 1996) . Correspondingly, if a patient is not willing to defer, that is because he does not regard the health worker’s understanding as a producer meaning of which he is a consumer. The underlying theoretical point about shared concepts applies in both kinds of cases: health workers and patients must be willing to defer to the same use of health expressions in order to associate the expressions with the same concepts.
Using disease as the main example, the article argues that if patients were normally willing to defer to a standard professional definition of “disease”, our common language concept of disease would be the expert concept with a meaning produced in the medical community. However, it is well documented that many patients are not willing to defer to health workers’ use of the word “disease”. In the light of Wittgenstein’s (1953) theory of contextual meaning, I will argue that any professional concept of disease—even a concept shared by many professionals—is only one among several concepts of disease. The fact that many patients think they are entitled to have their own understanding of “disease” constitutes a strong philosophical argument for conceptual pluralism—that there is more than one concept of disease within our linguistic community, and that none of these concepts constitute a general normative standard. This pluralism, I will conclude, is inconsistent with assumptions many theorists have made when they have analysed concepts of ill health.
The next section of the article outlines background assumptions about concept possession and partial understanding. Sections three and four use these assumptions to present and clarify the linguistic producer-consumer distinction and to argue that there are a variety of equally sound concepts of disease in our common language. The final part of the article discusses the practical implications of this conceptual pluralism by linking it to health discourse in general and doctor-patient interaction in particular.
Philosophical discussions of communication and language mastery are often based on the assumption that we do not have a complete understanding of a wide range of technical and theoretical words we use (Burge, 1979; Evans, 1982; Wright, 1984; Goldberg & Pessin, 1996) . The fact that we do not have an expert understanding of many terms in our common language, has led many theorists to focus on the relation between a person’s understanding of an expression and the concept he expresses (Peacocke, 1992; Burge, 2013) . The key philosophical issue can be phrased as a question about concept possession: how well must a person understand an expression in order to possess the expert concept that constitutes the normative meaning of the term, if the expression has such an expert concept?
In philosophy of mind and language, Burge’s (1979) “arthritis” case is famous and often discussed as a case of reference in debates about the relation between language mastery and concept possession. In this case, Burge imagines a patient who goes to his doctor complaining about pain and stiffness in his thighs, which he thinks has spread from his arthritis in the knee joints. He describes the case as follows:
…he thinks falsely that he has developed arthritis in the thigh. Generally competent in English, rational and intelligent, the patient reports to his doctor his fear that his arthritis has now lodged in his thigh. The doctor replies by telling him that this cannot be so, since arthritis is specifically an inflammation of joints (Burge, 1979: p. 79) .
Burge’s point here is that when the patient says that he thinks his arthritis has spread to his thigh, then he has made a mistake about how it is possible to have arthritis (the disease can only occur in joints). Given that the patient uses the term “arthritis” to express his belief, this is a mistake about the meaning of this term. Nevertheless, the patient has, as Burge (1979: p. 79) describes the case, some correct beliefs:
For example, he thinks (correctly) that he has had arthritis for years, that his arthritis in his wrists and fingers is more painful than his arthritis in his ankles, that it is better to have arthritis than cancer of the liver, that stiffening joints is a symptom of arthritis, that certain sorts of aches are characteristic of arthritis, that there are various kinds of arthritis, and so forth. In short, he has a wide range of such beliefs.
Understood like this, the patient does not have an understanding of “arthritis” that is radically wrong: He has many true beliefs, so he does not have a total misunderstanding. The question is how “far away” he is from a competent expert about the meaning of the work. In the light of the fact that the patient has the radically false belief about “arthritis” as applying to the condition in his thigh: does he qualify as someone who has the normative concept that does not apply to any condition outside joints, or is his understanding too weak? If he qualifies, as Burge argues that he does, then it is correct to ascribe to him beliefs involving the standard concept (so that his belief that his arthritis has spread to his thigh is false). If his understanding is too weak, then he expresses his own idiosyncratic concept that corresponds to his misconception (so his belief that his arthritis has spread to his thigh is true).
As numerous commentators of Burge’s “arthritis” example have noted, the question of which concept the patient possesses is a question of what he means when using “arthritis” (Woodfield, 1982; Bach, 1984; Crane, 1991; Travis, 1995; Nordby, 2004) . If a person means arthritis by “arthritis”, then he expresses the medical concept arthritis, as is true with all other concepts. For instance, if a person expresses the concept dog by “dog”, then he means dog by “dog” and his concept applies to nothing but dogs. If the person does not mean dog, then he expresses an alternative concept that matches his incomplete or incorrect understanding—a concept that applies to the meaning of the word in his personal idiolect.
In general, concepts we express relay the mental meaning of the word we use (Kripke, 1982; Davies, 1995; Peacocke, 2008; Martinich & Sosa, 2013) . As Davies (1995: p. 295) notes: “A concept is a way of thinking of an object or a property”. Our concepts can refer to objects but also states and events. The reference can be part of the external world but also something in our inner lives, like an experience of pain.
3. Communication and the Diversity of Understanding
The philosophical debate of shared understanding and concept possession is significant in analyses of patient communication. The crucial question is to what extent a health worker and a patient can understand a health term differently, but nevertheless associate the term with the same concept.
Note that this leaves open the question of how a correct understanding should be interpreted. In some cases patients have a more competent understanding than health workers they encounter. This can happen, for instance, when a patient with a rare chronic disease encounters a general practitioner who does not have detailed knowledge of the patient’s disease. Nevertheless, it is health workers who normally have the best understanding of health terms that are used in patient dialogue. Thus, in most cases, the philosophical question about concept possession is this: how well must patients’ understanding approximate health workers’ understanding, in order for patients to possess the same concepts as health workers?
In general, although many patients achieve substantial knowledge of their conditions of ill health, there is ordinarily an asymmetry of meaning in health worker-patient relations (Kreuter & McClure, 2004; Schiavo, 2007; Betsch, 2015; Berry & Yuill, 2016; Pagano, 2017) . Health professionals normally have a good understanding of relevant health vocabulary in their area of competence, while patients are typically laypersons (Kreps & Kunimoto, 1994; Helman, 2007; Nettleton, 2013) . Burge would say that the patient in the “arthritis” case has a partial understanding compared to the expert conception. His general idea is that a minimal understanding can be sufficient for possessing a concept:
Crudely put, wherever the subject has attained a certain competence in large relevant parts of his language and has (implicitly) assumed a certain general commitment or responsibility to the communal conventions governing the language’s symbols, the expressions the subject uses take on a certain inertia in determining attributions of mental content to him. In particular, the expressions the subject uses sometimes provide the content of his mental states and events even though he only partially understands, or even misunderstands, some of them (Burge, 1979: p. 114) .
For Burge, the “arthritis” case is just an example that he uses to illustrate this general view. The idea about partial meaning as a sufficient condition for concept possession applies to all discourse, and thus also to patient communication. Furthermore, in this area it has a striking application, in light of the general asymmetry in understanding of health terminology.
3.1. Philosophical Influence
Burge’s analyses of partial understanding and concept possession have been very influential in modern philosophy of mind and language (Goldberg & Pessin, 1996; Hahn & Ramberg, 2003; Peacocke, 2008) . The main reason many have thought that his arguments are convincing is that the arguments leave room for communication between experts and laypeople. If experts had their own specialized concepts, they would seldom be able to exchange concepts with laypeople.
In fact, the problem is deeper. Someone who holds that experts have their own concepts corresponding to their expert understanding must also, apparently, assume that two persons never have the same concepts. For if one does not make this assumption about idiosyncratic concept possession, then it is perfectly possible to accept that experts can exchange concepts with people who do not have an expert understanding. However, the problem with the idiosyncratic assumption is obvious; if we accept that every person has a concept that corresponds to his idiosyncratic understanding, then we must also accept the counterintuitive consequence that people never manage to exchange beliefs, thoughts and other states involving the same concepts. After all, two persons never understand a language in exactly the same way—they never have exactly the same beliefs about expressions of meaning.
This means that there is very good reason to accept, as Burge does, that people can understand an expression differently, but still associate the expression with the same concept. Some critics of Burge have nevertheless argued that Burge’s condition for concept possession is too modest; that it is counterintuitive to accept that someone who is so far away from a medical expert as the “arthritis” patient still possesses the same concept as the expert. It has been argued that more than a minimal understanding is needed to possess expert concepts (Woodfield, 1982; Bach, 1984; Jacobs, 1987; Crane, 1991; Travis, 1995) .
The critics have also argued that we do not have to reject the idea that experts and laypeople are able to exchange concepts if we accept that it takes more than a minimal understanding to possess expert concepts. If a normal, lay conception of arthritis is sufficient for possessing the expert concept of arthritis, then most patients are able to communicate with health workers about arthritis. The same holds with other medical terms and patient communication in general. If we accept that a normal lay understanding is sufficient for possessing the corresponding medical concepts, then patients normally possess the same concepts as health workers they encounter.
3.2. Methodological Challenges
The theoretical distinction between modest and demanding conditions of concept possession has implications for how one should analyse patient communication involving medical concepts. If the conditions are modest, health worker-patient communication is often successful in the sense that patients often exchange concepts with health workers. If the conditions for possessing medical concepts are demanding, then communication is not so often successful.
This theoretical difference constitutes a methodological challenge for everyone who aims to study health worker-patient interaction empirically. To determine whether communication of a concept is successful in a given type of relation is not merely an empirical issue; one cannot simply observe this (Alvesson & Schöldberg, 2010) . It is necessary to apply a theoretical framework about what it takes to exchange thoughts and beliefs involving concepts. But the choice between modest and demanding theories of concept possession is not simply a choice one is free to make. It is necessary to go into the various philosophical arguments for and against the conflicting perspectives. However, this is very much a disputed philosophical area, and if an empirical analysis rests on disputed arguments, then the analysis itself becomes disputed.
Further on in this article I will avoid this methodological problem by relying on an assumption that is accepted both by Burge and proponents of more demanding theories of concept possession, namely that deference-willingness to normative meaning is a necessary condition for possessing expert concepts (Peacocke, 1992) . This is an assumption that is crucial in any theory that assumes that the threshold condition for possessing concepts is weaker than the condition one has to meet in order to have an expert understanding.
Burge makes this assumption explicit in the arthritis case, by emphasizing that the patient defers: “The patient is surprised, but relinquishes his view and goes on to ask what might be wrong with his thigh” (Burge, 1979: p. 79) . This is crucial for making sure that the patient falls under the idea that “Global coherence and responsibility seem sometimes to override localized incompetence” (Burge, 1979: p. 114) .
Within this framework, we should, as Peacocke (1992: p. 29) observes, distinguish between a speaker “who belongs to our community, and a speaker who uses an expression in his own individual sense”. Thus, deference to normative meaning will be conceived to be essential for any theorist who accepts that individual scan possess expert concepts without having an expert understanding. As long as a person does not have a complete understanding, he needs to be willing to defer to persons he regards as experts.
4. The Producer-Consumer Distinction
One might think that the condition of deference does not make a difference with respect to communication. Being deference-willing is an attitude, and how can a person’s attitudes determine the content of his concepts? The key to understanding how this is possible is to think of laypersons who defer to experts’ meaning explanations as consumers of the experts’ producer understanding.
In philosophy of mind and language, this producer-consumer distinction is often associated with Hilary Putnam’s theory of degrees of linguistic competence (Putnam, 1975, 1986; Evans, 1982; Goldberg & Pessin, 1996) . For Putnam, persons who have an incomplete understanding of an expression can nevertheless use it in accordance with its general meaning. Using his own incomplete understanding of “elm” as an example, Putnam (1996: p. 12) argues that “the extension of ‘elm’ in my idiolect is the same as the extension of ‘elm’ in anyone else’s, viz., the set of all elm trees”. Putnam argues that even though he cannot “tell an elm from a beech tree”, he means the same as an expert when he is using “elm”: The reason is that he thinks of the expert’s producer understanding as correct not only for the expert, but also for himself as a consumer.
According to Putnam, the overall role expressions have in a linguistic community is formed by both consumers and producers. It is their joint effort that makes up the general uses of a term, and thereby its overall meaning. Putnam (1996: p. 14) calls this joint enterprise a division of linguistic labour:
Every linguistic community… possesses at least some terms whose associated criteria are known only to a subset of the speakers who acquire the terms, and whose use by the other speakers depends on a structured cooperation between them and the speakers in the relevant subsets.
To illustrate, Putnam (1996: p. 14) uses a metaphorical distinction between two kinds of tools: “There are tools like a hammer or a screwdriver which can be used by one person; and there are tools like a steamship which require the cooperative activity of a number of persons to use. Words have been thought of too much on the model on the first sort of tool”. According to Putnam, the overall meaning of a word in a linguistic community, as opposed to an idiosyncratic meaning for an individual or a group of speakers, involves much more than producers’ understanding. It also involves all the thoughts and beliefs consumers have about the meaning of the word.
4.1. Health Terms
Putnam’s analysis of the division of linguistic labour is completely general, but it has a striking implication in the area of health discourse. For just like “water” and “elm”, many medical terms have a producer meaning of which many patients do not have a complete understanding. Furthermore, confronted with health workers’ explanations of medical terms, patients are normally willing to defer to these explanations, even when their own understanding is partially incorrect. As Gillon (2001: p. 508) observes,
...even common words such as “cancer” are likely to be radically misunderstood by patients unless they have had a medical training. The wide range of conditions and prognoses and all other technical nuances implied by the word are probably not taken into consideration and are often replaced by a single dark understanding.
Obviously, when patients realize that they have made factual mistakes about the meaning of a natural kind word like “cancer”, they typically adjust their understanding so that it better approximates health workers’ understanding. “Cancer” is just one example of a medical producer term of which patients conceive themselves to be consumers. Patients believe that they are developed within the expert profession.
Other producer expressions in health discourse include common terms that refer to states of disease like “cancer”, “aids” and “diabetes”, less well-known expressions like “multiple sclerosis”, “spinal stenosis” and “obstructive pulmonary disease”, and more technical denotations like “computed tomography”, “chemotherapy”, “stent insertion”, “intubation” and “electrocardiogram”. There are countless expressions that have a normative meaning in the medical profession. In general, when patients are confronted with explanations that refer to treatment, procedures or somatic conditions of ill health, they typically defer to the explanations because they think the correct meaning can be found in the professional discourse.
The difference between such producer terms and subjective health expressions is striking. Patients do not think of disputed expressions like “disease”, “illness” and “sickness” as expressions that belong to the medical profession. What is special about these expressions is that it has been extensively documented that patients’ understanding is heavily shaped by socio-cultural frameworks of meaning, and that patients think they are entitled to use the expressions in accordance with these frameworks (Galanti, 1991; Helman, 2007; Burnard & Gill, 2008; Nettleton, 2013; Berry & Yuill, 2016; Spector, 2017) . The meaning of the expressions is conceived to be grounded in these frameworks (Kreps & Kunimoto, 1994; Snehendu et al., 2001; Macklin, 2006; Weitz, 2013) . Consequently, patients do not think of themselves as consumers of an external expert understanding.
Note the difference to Burge “arthritis” case as described above. When corrected with his mistake, the patient “relinquished his views”. He did not understand himself as an authority on the meaning of “arthritis”, and he did not think he was entitled to have a personal understanding that deviated from the medical meaning. He deferred to the doctor’s use, not because he thought that the doctor’s use matched his understanding, but because he conceived the doctor’s use to be the correct use in their common overall language. Contrary to subjective terms, “arthritis” had a general standard expert meaning for the patient.
4.2. Communicative Challenges
It is important to emphasize that the linguistic producer-consumer distinction does not imply that it is straightforward for health workers to create a platform of shared concepts when patient dialogue involves medical terms that refer to somatic conditions. The communicative challenges surrounding medical terms are profound (Kreps & Kunimoto, 1994; Helman, 2007; Schiavo, 2007; Dutta, 2008; Pagano, 2017) . Knowledge gaps, power relations, communicative noise and patients’ mental and physical states make it difficult to explain the meaning of medical expressions (Kazarin & Evans, 2001; Zoller & Dutta, 2008; Hogg & Holland, 2010; Wright et al., 2013) .
The challenges concerning subjective terms do not involve the same kind of epistemological problems, but the challenges are not any less intriguing. The reason is twofold: First, there is a greater diversity of norms of meaning of such terms in the lay population. Second, there is not even agreement about their exact meaning in the medical profession.
The fact that patients often stick to their own subjective understanding, demonstrates that it is especially difficult for health workers to create a shared understanding. In contrast to the use of medical terms, use of subjective terms does not involve a development towards a shared understanding as in the “cancer” case above. Consider the following case:
A patient suffers from fatigue, dizziness and a mild form of depression. He has been assessed for a variety of diagnoses but no objective findings have been made. His doctor tells him that no findings can establish that the patient suffers from a disease. The patient responds that “surely I must have a disease with all these problems”. The doctor explains that this cannot be documented as long as no findings have been made. The patient understands the doctor’s thinking, but he thinks to himself: “The very fact that I have all these problems means that I must have a disease”.
In this case the linguistic society as a whole disappears as the unity of reference for the patient’s concept. It is not possible to understand the extension of the patient’s concept by elucidating a medical “expert” meaning of illness. In order to understand the patient’s concept it is necessary to identify an alternative framework of reference—the framework that the patient thinks entitles him to reject the doctor’s explanations. This framework is grounded in a socio-cultural setting that is different from the doctor’s understanding.
Thus, for the purposes of comparing the doctor’s norms of meaning with the patient’s norms of meaning, it is necessary to compare two contexts of understanding in our linguistic community. In such comparative analyses of contextual meaning, both a horizontal and a vertical level of inquiry are relevant. The aim of horizontal analysis is descriptive—it is to understand the similarities and differences between various contexts of understanding—how they overlap and how they are strikingly different. The aim of the vertical analysis is normative—it is to understand if some contexts of understanding have a higher status than others in the sense that they provide norms for the correct understanding.
The latter normative question about correctness is philosophical and not empirical. It cannot be answered simply by studying actual use of language. It must instead be addressed in the light of theories of meaning. Furthermore, the significance of such theoretical answers is obvious. It is a widespread view that the question of how basic health concepts should be understood has huge epistemic, practical and economic implications (Albert et al., 1988; Nordenfelt & Twaddle, 1993; Hofmann, 2001, 2002) .
5. Wittgenstein on Meaning
The explanatory limits of the producer/consumer distinction when applied to basic health terms can be interpreted from a variety of conceptual frameworks. Giving priority to one perspective would require a special justification. A more reasonable assumption, at least prima facie, is that different analytical frameworks can jointly contribute to widen our understanding of incompatible frame works of meaning in health discourse.
In modern philosophy of mind and language, one of the most central analyses of the diversity of language meaning is Wittgenstein’s theory of language mastery, presented in his famous Philosophical Investigations (1953). In this section I will show how Wittgenstein’s concept of a language game can give us a deeper understanding of why not all health communication conforms to the producer-consumer distinction as described above. This understanding, I will go on to argue, can explain why basic health concepts like disease cannot be analysed in the way many theorists in the philosophy of medicine and health care have tried to analyse them.
For Wittgenstein, the way we understand expressions must fundamentally be derived from how we use them: “The meaning of a word is its use in language” (Wittgenstein, 1953: p. 20) . Wittgenstein contrasts this approach to understanding with the idea that we have introspective access to inner ideas that can tell us how we should use language. Wittgenstein argues that if we had such inner ideas, then they would have to be rules that guide us in our application of words—they would have to instruct us how to use words in all kinds of contexts. But this is utopic: a rule formulation must be extremely complex to cover all the various contexts in which an expression applies (Cavell, 1979; Kripke, 1982; Stroud, 1996) .
However, the main problem, Wittgenstein argues, is even more fundamental: rule formulations must be interpreted—they must be ascribed meaning—but that requires new explanatory formulations that in themselves must be interpreted. Any attempt to introduce a rule for interpreting a rule will simply rephrase the problem. A rule understood as a grammatical expression does not have an action-guiding content. The surface expression can be interpreted in all kinds of directions: “This was our paradox: no course of action can be made out to accord with a rule, because every course of can be made out to accord with the rule” (Wittgenstein, 1953: p. 81) .
Consider as an example the expression “S”, and a rule formulation “X” considered to be a candidate for an explanation of the meaning of “S”. As long as “X” is a grammatical formulation its meaning must also be explained, and we can attempt to do this by introducing a new rule formulation “Y”. But then the problem remerges: the expression “Y” must also be explained, and it is necessary to introduce yet a new formulation “Z”. The only way to avoid the regress is to assume that we know the meaning of a rule formulation. But insofar as the aim is to explain what the meaning of a formulation is, we cannot take this for granted in the explanation (Cavell, 1979; Kripke, 1982; Harris, 1988; McDowell, 1998) .
5.1. The Solution
Wittgenstein’s solution to the problem of rule-following is to argue that it is wrong to analyse language mastery from a first-person perspective from which all thinkers are guided by instructions of use to which they have introspective access. According to Wittgenstein, we must instead analyse meaning from a third person perspective, as socio-cultural practice that establishes what we mean (Wittgenstein, 1953; Kripke, 1982; Harris, 1988; McDowell, 1998; Horn, 2005) .
For Wittgenstein, this “outside” perspective implies that mastering a language expression is a matter of being be able to master it in accordance with contextual rules. Sometimes speakers conceive such contexts to be very wide, so that they include the whole community. This holistic idea of a context is emphasised by Kripke (1982) in his influential Wittgenstein interpretation; he uses “plus” as example and outlines Wittgenstein’s core point as the view that
…any individual who claims to have mastered the concept of addition will be judged by the community to have done so if his particular responses agree with those of the community in enough cases, especially the simple ones … An individual who passes such tests is admitted into the community as a normal speaker of the language and member of the community (Kripke, 1982: p. 92) .
Understood this way, there is an obvious link between Wittgenstein’s contextual analysis of meaning and the producer-consumer distinction as explained above: to observe that a person conforms to uses of “plus” in a limited number of situations is not the same as seeing that the person conforms to all the norms for using “plus” in our community. Within Wittgenstein’s analytical framework, the person’s conception of himself as a member of the linguistic community becomes crucial. The person means plus if he wants to understand “plus” in accordance with standard use, and consequently defers to this in future situations. In this case the whole society is the unity of reference. As long as he is deference-willing, he is a consumer of the general meaning and it is correct to ascribe to him the concept plus.
However, Wittgenstein’s philosophy can also be used to explain why the producer/consumer distinction does not always apply, for we might also imagine someone who does not defer to the standard “expert” use of an expression. He might use it in his own contexts, but not defer when he is confronted with the general use. If so, then he has his own standards, and he is not a consumer of the public meaning. As Kripke (1982: p. 92) observes, those “who deviate are corrected and told (usually as children) that they have not grasped the concept of addition”. Kripke outlines Wittgenstein’s general idea as follows:
One who is an incorrigible deviant in enough respects simply cannot participate in the life of the community and communication… A deviant individual whose responses do not accord with those of the community in enough cases will not be judged, by the community, to be following its rules (Kripke, 1982: p. 92) .
Furthermore, for many expressions it is not so clear how the public meaning should be defined, so that there is no clear producer meaning that constitutes a condition for deference-willingness. This was the case in the abovementioned example involving “disease”. There is no commonly accepted definition of disease equivalent to the specific norms for the use of “plus”.
5.2. Language Games
For Wittgenstein, the diversity of understanding can be explained under the notion of a language game—a systematic use of language that is governed by explicit and implicit rules. Some language games represent what most speakers would regard as a general meaning in the community, as in the “plus” example above. However, Wittgenstein stresses that language games can be local as well. In such games, speakers tend to stick to the understanding they have when confronted with other language games (Harris, 1988; Mcdowell, 1998; Horn, 2005; Salehi-Nejad, 2014) .
In philosophical debates, it is especially these local language games that have received attention. The idea of a limited practice of meaning has often been used as a key concept in critical arguments to general theories of meaning that do not incorporate the diversity of understanding of disputed terms (Cavell, 1979; Harris, 1988; Lawn, 2004) . Furthermore, it is widely recognized that local language games do not fall under the general producer/consumer distinction. Speakers in such games think they are entitled to their own standards, but the source of this entitlement is not the whole linguistic society. They believe that their understanding is grounded in a form of contextual social capital that justifies their use.
As long as speakers of local language games have this experienced entitlement, they will not defer to norms of meaning in other language games. If they meet someone who regards himself as an expert about meaning, they will defer only if they think the person’s “expert” understanding represents an objective standard compared to their own. It is precisely this condition that was not met in the disease example above: the patient did not think that the doctor’s understanding was a legitimate norm in the patient’s own framework of meaning.
6. A Pluralistic Perspective on Health Discourse
The relevance of Wittgenstein’s analysis of language mastery in the area of health discourse is striking. According to Wittgenstein, in order for two persons to express the same concept by a word like “disease”, they must be willing to defer to the same norms of meaning. But “disease” is a word that is used in many different language games, and speakers of one language game are often not willing to defer to the rules of other language games.
Within such a pluralistic perspective on meaning, it is still possible to do conceptual analyses by clarifying particular concepts of disease, as they are understood within language games. Doing this is the same as clarifying the rules of specific discourses about disease. These kinds of conceptual analyses of “local” meaning are, in fact, often developed in sociology of health and medical anthology, and they can yield valuable insights about contextual discourse of ill health (Foster & Anderson, 1978; Macklin, 2006; Burnard & Gill, 2008; Weitz, 2013) . Wittgenstein’s philosophy is not only compatible with such analyses; his description of language games offers a theoretical foundation for them. In fact, Wittgenstein’s philosophical terminology can be used as conceptual tools in any attempt to clarify the diversity of language meaning (Cavell, 1979; Harris, 1988; Lawn, 2004; Salehi-Nejad, 2014) .
Understood like this, Wittgenstein is not opposed to the idea of conceptual analysis as such. What W is incompatible with is the idea that basic concepts of disease have one general meaning that can be analysed, or one meaning that should prevail, when there are conflicts between language games. In such cases, Wittgenstein argues, the overall meaning of a concept is pluralistic—there is not one meaning that is more correct than others.
This kind of meaning pluralism should not be conflated with the claim that language games are isolated and incommensurable entities that are not influenced by each other. Furthermore, it would not be in the spirit of Wittgenstein’s philosophy to argue that health discourses have stable and fixed structures. Wittgenstein stresses that language games are dynamic, fluid and overlapping. In his analysis of contextual meaning Wittgenstein focuses heavily on how people revise their understanding, how frameworks of meaning change, and how language games sometimes melt together.
Conceptions of disease can change in all these ways. Not only patients, but also health workers, can revise their ideas of what it means to have a disease. How this actually happens is very much an empirical question. Wittgenstein’s point is that real change in a person’s understanding must come from within: a new meaning explanation has to be experienced as correct from the person’s perspective, no matter what the person’s professional or lay framework of understanding is. Thus, Wittgenstein is not only opposed to the claim that there are outside standards that can capture a core meaning of terms that are used in many different language games. He is also opposed to the idea that there can be outside expert standards that speakers should conform to even though the standards do not match their understanding. Speakers should be allowed to use their own frameworks of meaning as norms for evaluating alternative frameworks.1
6.1. Conceptual Analyses
We are now in a position to understand, philosophically, why the diversity of understanding of basic health terms constitutes a fundamental obstacle to the project of developing general analyses of health terms. Traditionally, analyses of concepts like disease, illness and sickness have sought to capture the common meaning of the concepts within our common language. Hartman formulates the general idea of conceptual analysis in an illuminating way:
Typically, attempts at philosophical analysis proceed by the formulation of one or more tentative analyses and then the consideration of test cases. If exactly one of the proposed analyses does not conflict with “intuitions” about any test cases, it is taken to be at least tentatively confirmed. Further research then uncovers new test cases in which intuitions conflict with the analysis. The analysis is then modified or replaced by a completely different one, which is in turn tested against imagined cases, and so on (Harman, 1999: p. 139) .
In a comparative test such as this, the concept of a shared language is essential. The traditional assumption is that speakers of the language share a set of intuitions about what a language expression means. It is these intuitions that are supposed to tell us how a concept analysis applies. They are the criteria that any proposed definition must match.
When applied to lay health terms, the problem is obvious. As long as the meaning is pluralistic within the overall community, it is impossible to extract one meaning that captures the variety of uses in our common language. This is a problem that confronts any analysis that aims to define disputed concepts of disease in general, no matter how the definition seeks to define the concepts.
This problem has a wide scope. In the last decades, there have been a multitude of attempts to develop general definitions of basic health terms. As conceptual approaches, they can be divided into two types (Nordenfelt & Twaddle, 1993; Hofmann, 2002) . One type is naturalistic definitions—definitions that seek to explain disease in naturalistic terms. There is no commonly accepted definition of what a naturalistic term is, but the guiding rule has been that a term is naturalistic if and only if it is used in explanations in the natural sciences. Perhaps the most influential naturalistic definition is Boorse’s (1997: pp. 8-9) suggestion that “disease as a type of internal state which is either an impairment of normal functional ability, i.e. a reduction of one or more functional abilities below typical efficiency, or a limitation on functional ability caused by environmental agents”. Other naturalistic definitions have focused more on biological descriptions, like Marinker’s view that disease is a “pathological process most often physical as in throat infection, a cancer of the bronchus, sometimes undetermined in origin, as in schizophrenia” (Marinker, 1975: p. 81) . There are many variations of naturalistic definitions, but they have all attempted to capture a common meaning of “disease”.
The same assumption about generality unites evaluative definitions. Such definitions have attempted to clarify disease partly or wholly in evaluative terms, by appealing to the evaluative judgments a person or group of persons would make involving the concept (Nordenfelt & Twaddle, 1993; Hofmann, 2002) . A classical example is Scadding’s definition that a disease is “the sum of the abnormal phenomena displayed by a group of living organisms in association with a specified…set of characteristics by which they differ from the norm for their species” (Scadding, 1967: p. 25) . The details of the evaluative definitions differ, but they all attempt to define disease by capturing a common meaning in normative vocabulary. The problem, however, is that there is no such meaning.
6.2. Possible Responses
Is there any way the traditional conceptual analysists can avoid this problem? One way would be to redefine the scope of the analysis to a narrow understanding in one specific language game. The idea would be that it is possible to arrive at a substantial analysis of a concept if we stick to one determinate area of discourse. The overall analysis would then be stipulative—its correctness would only depend on whether it matches the defined area of discourse and not the language as a whole.
The problem with this strategy is that if the aim of the conceptual analysis is to capture a narrow understanding, then the normativity of the analysis becomes very limited. The analysis cannot have general normative consequences. The reason, as Nordenfelt (1987: p. 8) argues, is that an analysis that does not capture a common understanding “would not be used in ordinary discourse, and would therefore be of no interest to us”. Those who understand “disease” in ways that do not fit the target understanding of the conceptual analysis could simply say that they have chosen to understand the word using other language games. They would, in Christopher Peacocke’s words, “be a group of speakers who choose to use a word in their own individual sense” (Peacocke, 1992: p. 29) . Thus, the narrow stipulative analysis could not function as a standard that could be a universal platform of reference in health discourse. But establishing such a platform has precisely been the aim of the conceptual analyses.
A second way of trying to avoid the pluralistic objection to traditional conceptual analyses, would be to argue that philosophical theories of conception possession and language mastery are irrelevant for the purpose of defining controversial concepts. According to this strategy, definitions of the nature of basic health concepts like disease and illness should not pay attention to controversial philosophical theories about concepts, language games and conceptual analyses: if the analysis rests on controversial premises, then the analysis itself becomes controversial.
The problem with this strategy is that the challenges about the diversity of understanding are not theoretical challenges. It would be a misunderstanding to think that the significance of the division of linguistic labour depends on philosophical theories. The facts about the diversity of understanding are empirical facts that any analysis must be compatible with.
The third possible way of responding to the pluralistic analysis is to argue that the fact that speakers understand “disease” in various ways is no guarantee that one cannot find a common core of meaning in various language games. According to this suggestion, we should continue searching for an understanding that involves laypeople, many health professions (in a wide sense) and patients’ understanding.
There are two problems with this strategy. First, a meaning explanation of a term is not a full-blooded definition if it only captures an aspect of the meaning of the term. Such an explanation would only describe a necessary condition of meaning. Consider as an analogy how being a mammal is a necessary condition for being a dog, but is not a definition of a dog.
Second, it seems utopic to think that it is possible to find a substantial common element in the various conceptions of disease. We should not rule out a priori that it is possible. But in the light of the fact that no theorist has managed to formulate a definition that has received widespread acceptance, it seems highly implausible to assume that there is a “hidden” meaning that we have not yet found. This kind of skepticism is in the spirit of Wittgenstein’s (1953) later philosophy. His analyses of language games are not presented as a knock down argument against the possibility of general analyses. Wittgenstein is opposed to this kind of armchair metaphysics (Cavell, 1979; Harris, 1988; Lawn, 2004) . According to Wittgenstein, it is just ordinary facts about language that make it overwhelmingly natural to assume that subjective terms are related to each other through family resemblance—a network of meaning in which one use might be totally different from a use far away in the network. This diversity of understanding makes it overwhelmingly natural to assume that it is impossible to formulate general definitions of disputed concepts.
In sum, all attempts to develop a general definition of controversial concepts of disease face the same problem. The facts about the diversity of understanding strongly suggest that a definition cannot be formulated. There is no way around this problem. It is not based on idiosyncratic assumptions or controversial theories, only on the assumption that there is no univocal meaning to be found. This assumption is not a pirori correct—we have no guarantee that language games will not change. But appealing to this possibility would be to misunderstand the point: as long as empirical research continues to uncover a variety of incommensurable conceptions of disease, there is good reason to believe that the project of finding general definitions is on the wrong track.
7. Implications in Health Communication
To argue for a pluralistic analysis and point out that it is inconsistent with the idea of a general concept definition is of limited value if the analysis does not have practical implications. So what is the applied dimension of pluralism?
From a professional perspective, the most important implication is that health workers should acknowledge that patients are entitled to understand basic health terms in accordance with the rules of their own language games. There is no professional understanding that constitutes a normative standard: when health professionals’ understanding collides with patients’ understanding, no external “expert” criteria can be used as a basis for favouring the professional understanding. Granted, if there were such criteria, then it would be possible to use them to identify one language game, and possibly a professional language game, as the correct. But the pluralistic analysis implies that there are no such criteria.
Some might nonetheless think that this implication opens up for an unacceptable relativism. For how can health workers and patients understand each other if they do not have the same health concepts? Formulated as a reductio ad absurdum objection, some might argue that pluralism has the implausible consequence that health workers and patients seldom are able to communicate, and that pluralism therefore is implausible in and of itself, as a philosophical position.
How should the pluralist respond to this objection? He has to accept—as a key premise in his own position—that if health workers and patients are not willing to defer to the same norms of meaning, then they do not have a shared platform of concepts. He also has to accept what is well documented—that patients can misinterpret information health workers intend to convey, and fail to act in ways that they are supposed to act, if they fail to grasp the meaning of health workers’ language. Extensive empirical research has shown that poor communication related to incompatible conceptions of ill health can have a number of negative consequences (Schiavo, 2007; Wright et al., 2013; Jeffreys, 2016; Pagano, 2017 ).
This means that if pluralism implied that health workers and patients never shared concepts, the reductioad absurdum objection would be convincing. But this is only a conditional, and the pluralist can legitimately claim that the antecedent is false. The reason is that the disputed terms are only a small minority of health terms. Holding that pluralism is committed to a grand scepticism about successful exchange of health concepts rests on an incorrect picture of the prevalence of subjective terms. For most terms, the producer/consumer distinction applies: health workers and patients have a sufficiently similar understanding and they are willing to accept the same norms of meaning produced in the medical community. It is only use of terms that are grounded in entirely different language games that falls outside the scope of the producer/consumer distinction.
7.1. Finding Common Ground
The practical implication is obvious: for health professionals, a key to secure communication is to steer away from disputed terms. Consider again a situation in which a doctor has initiated discourse about disease by telling a patient that it is not documented that the label “disease” fits his condition. The patient reacts. He thinks that his symptoms are sufficient, and that the doctor means to question his status as a person who should receive help and treatment from public health services. The doctor, however, does not mean to create this kind of suspicions. He is merely concerned with the fact that there is no biomedical diagnosis that has been found to apply.
In this case, misinterpretation could have been avoided if the doctor had focused more directly on the patient’s condition by using producer-consumer expressions or everyday terms that speakers tend to understand in the same way. We can assume that neither the doctor nor the patient doubted that the patient told the truth about his condition. Their disagreement was related to how the condition should be labelled. If the doctor had not initiated talk about “disease”, but used less controversial language in the conversation, the patient would not have been so frustrated. He would not have thought that the doctor meant to question his status as a patient with serious health problems.
It is, obviously, sometimes difficult to avoid use of controversial terms, but in most cases it is possible to find fairly common ground. Even subjective terms often have a common core. “Pain” might normally be used in discourse about pain without much controversy—this term is ordinarily understood in much the same way. On the other hand, using “illness” when referring to pain risks creating substantial misinterpretation. It is, for instance, well documented that some patients have a holistic view on illness—they might think of illness as extending to existential illness like grief, depression, and lack of energy (Macklin, 2006; Nettleton, 2013) . Health professionals, on the other hand, might restrict the term illness to an experienced conscious with a qualitative content. Such incompatible frameworks can lead to substantial misunderstanding. A health worker who uses illness in a narrow sense can be interpreted from a holistic framework.
This does not mean that use of terms that are understood differently should be dropped altogether. Sometimes it is overwhelmingly natural to use terms that are understood in incompatible language games. This might typically happen when a patient initiates use of disputed terms. In such cases it is important to clarify and prevent potential conflicts of meaning. An example could be the tension between patients’ use of “sicknesses” and the idea of a sickness certificate. It has been well documented that patients understand “sickness” in a variety of ways, and that many of the associations are incompatible with the formal requirements for having a sickness certificate (Hofmann, 2002; Helman, 2007; Weitz, 2013) . Thus, the formal requirement scan be introduced as a specific meaning, and patients can be informed that they are not meant to incorporate all the possible associations patients might have about “sickness”.
More could be said about the importance of acknowledging different language games in health discourses, but that would fall outside the limits of this article. The aim has been to illustrate the explanatory limits of the producer-consumer distinction: it does not apply when terms are used in language games in which speakers are entitled to understand the terms in accordance with contextual rules. It is therefore imperative to respect a variety of uses of basic health terms and try to avoid disputed terms as well as realistically possible.
The question of the meaning of health and illness is fundamentally grounded in our use of language—how each and every one of us thinks that concepts of ill health apply. The fact that the issues concern us all, gives the philosophical analyses of the concepts a striking significance. Conceptual analyses focus on how we actually think, and how we should think, about the application of concepts (Martinich & Sosa, 2013) .
Drawing on Wittgenstein’s theory of language games, I have argued that a pluralistic analysis incorporates both these dimensions. Descriptively, the analysis captures the diversity of understanding of basic health terms. Normatively, it gives substance to the question of how health workers should secure communication involving disputed concepts. The most striking implication of the pluralistic analysis is that different language games should be respected. Contrary to what many theorists have assumed, for certain concepts, it is not possible to develop univocal definitions that health workers can use as normative standards in health discourse.
For participants in such discourse, and health professionals in particular, there are several ways of acknowledging this pluralism in communication practices. The most important point is to understand language games. By realizing how incompatible language games are different, health workers can uncover potential misunderstandings. Moreover, explaining differences of meaning to patients can be done without initiating a debate about the correct meaning of disputed terms. If both parties recognize the other’s understanding, then both parties understand what the other is talking about. This is necessary for getting into a position from which it is possible to have a real discussion about facts.
The necessary and sufficient condition establishing a communicative platform despite the diversity of understanding is to find a common language. It is a sound prima facie principle that health communication should not, as far as possible, involve the use of controversial concepts that health workers and patients understand differently. Of course, sometimes patients introduce disputed words and make them central in the dialogue. But health workers can try to move the focus away from such labels and focus more on mutually accepted concepts.
When it is impossible to overcome communicative barriers this way, a second option is to initiate meta conversation about meaning. By making it clear to patients that there is a conflict of language games, patients can understand that apparent real disagreement about facts is agreement about meaning. Furthermore, such an insight can lead patients to revise their own conceptions and, possibly, create an attitude of deference-willingness to health workers’ understanding, simply because patients think these language games are sensible frameworks of meaning.
In the final instance, this kind of normativity cuts both ways. Patients might defer to health workers, but health workers might also defer or adjust to patients’ perspectives when they think they are sensible. By remembering this principle about the possibility of natural deference from within, health workers can both accept the Wittgensteinian point about respecting different language games and fulfil the normative aim of creating a platform of shared concepts in patient communication.
 For Wittgenstein, this pluralism about meaning does not legitimate any kind of relativism about knowledge. The reason, he argues, is that questions of meaning are not subject to the same claims to reason as questions of truth and knowledge. If something is true then it is true, but if an expression means something for a person, then it does not have to mean the same for another person. This, for Wittgenstein, is the crucial difference between epistemology and philosophy of language.
 Betsch, C. (2015). Improving Medical Decision Making and Health Promotion through Culture-Sensitive Health Communication. Medical Decision Making, 36, 811-833.
 Burge, T. (1979). Individualism and the Mental. In P. French, T. Uehling, & W. Wettstein (Eds.), Midwest Studies in Philosophy (Vol. 4, pp. 73-121). New York: Wiley.
 Kazarin, S., & Evans, D. (2001). Health Psychology and Culture. Embracing the 21st Century. In S. Kazarin, & S. Evans (Eds.), Handbook for Cultural Psychology. Waltham, MA: Academic Press.
 Stroud, B. (1996). Mind Meaning and Practice. In H. Sluga, & D. Stern (Eds.), The Cambridge Companion to Wittgenstein. Cambridge: Cambridge University Press.
 Worrall, J., & Worrall, J. (2001). Defining Disease: Much Ado about Nothing? In A. Tymieniecka, & E. A. Gazzi (Eds.), Analecta Husserliana LXXII (pp. 33-55). Dordrecht: Springer.