Introduction

Studying scientific arguments when they are fully fledged, fully packaged, and fully ornamented may be unnecessarily cumbersome for science studies researchers in general and for philosophers of science in particular. In these arguments, many aspects may remain unverifiable, since we only examine their rationality or normativity as detached from their foundations (i.e., how the related experiments happened in reality). Moreover, scientists are brilliant at making their scientific findings appeal to minds (Kekes, 1995). They are proficient in burying and hiding their flaws, mistakes, or imperfections (Latour, 1987). More interestingly, scientific arguments are not necessarily stated or explicit. They may be unstated or unverbalised, but still govern the scientists’ choices and decisions (Polanyi, 1958; Collins, 1974; MacKenzie and Spinardi, 1995; Pinch et al., 1996; Collins, 2001; Collins and Weinel, 2011). On the one hand, tacit knowledge underlying the scientists’ decisions within research labs makes a substantial part of scientific knowledge (Collins, 2001). On the other hand, it is hard to formalize (Polanyi, 1983; Collins, 2001), as it is acquired through undocumented practices and interactions (Sternberg, 1999; Nonaka and Takeuchi, 1995). We, therefore, favor studying scientific arguments in situ and in-the-making to be able to spot tacit knowledge, as an entry point. Targeting tacit knowledge in situ enables social sciences researchers to identify what we here call unstated arguments. An unstated argument, as we conceive of it, is linguistically less than an argument, as it has never been verbalized. However, it is more than tacit knowledge, as it may be a complete argument, except that it is expressed nonverbally (neither orally nor written, but through practice).

To be able to identify unstated arguments, we start from a specific tacit knowledge as demonstrated in a specific observable scientific practice and attempt to figure out how knowledge sharing works in a rather inductive way. Collins (2001) defines tacit knowledge as “knowledge or abilities that can be passed between scientists by personal contact but cannot be or have not been set out or passed on in formulae, diagrams, or verbal descriptions and instructions for action” (p. 71). We, therefore, identify the unstated scientific argument as a type of argument in which the conclusion or main claim is not explicitly stated. Instead, the conclusion is implied. Taking a descriptive perspective, an unstated scientific argument can be identified if: (1) we see a claim as practiced nonverbally (as an observed practice) through tacit knowledge; (2) we solicit evidence from the scientists when we as social scientists deliberately bring up a conversation with them about that specific nonverbal claim; and (3) a counterevidence is provoked to examine the basis of the evidence itself. We would therefore categorize that specific tacit knowledge as an unstated argument. To put that into practice, we opt for ethnography, as it allows social scientists to study other sciences in their own habitat (i.e., research labs) and look into science prior to packaging. Unlike unstated arguments, verbalized scientific arguments are abundant in the literature. Our literature review in the next section focuses on how studies on scientific arguments have ramified epistemologically. We favor Walton’s perspective on the argument, which “departs from the impersonal framework of deductive logic, [and sees] an argument as a dialog exchange between two parties who are reasoning together. No longer exclusively concerned with deductive and inductive forms of argument, argumentation theory considered many forms of presumptive inference based on intelligent guessing that leads to a tentatively acceptable conclusion on one side of a dialog.” (Walton, 1998, p. 72).

In the present study, a cancer research lab in Brussels is targeted. We therein examine scientific arguments inductively rather than deductively, descriptively rather than prescriptively, and interpretatively rather than from a positivist perspective. Firstly, we conduct an ethnographic observation. By the same token, we spot and identify tacit knowledge. Discussions with cancer-research scientists enable us to zoom in on a specific scientific practice. From there, we discern the claim of the unstated arguments (a claim expressed nonverbally through the scientists’ practices), which we then subject to explicitation through tape-recorded open interviews with the scientists (our participants). We finally end up with a verbalized argument (a verbal claim, evidence, and counterevidence), which we analyze using Discursive Psychology as a theory and method of discourse analysis.

The rhetorical tradition in scientific argumentation studies

The rhetorical tradition spanning from ancient Greece to contemporary discourse theory in the United States and Europe addresses a wide range of descriptive, explanatory, and prescriptive topics. Some of these topics are concerned with the speaker’s “invention”: the discovery of arguments, which is what we aim at in the empirical part of this study (i.e., digging down for the unstated arguments), while others are concerned with the texts criticism. In the rhetoric of science, two approaches, both based in U.S. universities, are influential: Speech Communication and English Composition (Keith and Rehg, 2008).

The Speech Communication Approach theorists use Aristotle’s political speech predicament as a major topic of discussion and place emphasis on group discussions aimed at choosing a course of action rather than on dialectical dialogs designed to prove or disprove theses. Within the English Composition approach, an argument is seen as one of the modes of discourse and treated as a product (a linguistic product or performance mirroring the learners’ grammatical, semantic, and logical reasoning performances). In both approaches, rhetorical theorists do not separate the internal component of reason from its external context (Keith and Rehg, 2008). Interestingly, in our present case study, we start from the external context (the cancer research lab in which the argument was born through tacit knowledge) and then examine the argument itself (internal reasoning).

Around the 1970s, argumentation studies shifted the interest to speech and discourse, questions about the use of instruments or ideas in a particular location, knowledge production and intervention, and science as a mode of working rather than a system of propositions. Diverse studies have provided a realistic view of what scientists actually do, and we wish to situate the present study within this perspective. Some studies were inspired by Kuhn’s interpretation of paradigms as exemplars, others by Polanyi’s idea of tacit knowledge, or by Wittgenstein’s attention to forms of life. Some adopted the new approach because of a commitment to ethnomethodology, and others turned to anthropology and its ethnographic methods (Amsterdamska, 2008). Some studies highlight the mediating role of instruments and technologies. Others draw attention to their restraints in knowledge production, tacit knowledge transfer, and standardization efforts to reduce uncertainty or facilitate communication (Burri and Dumit, 2008).

Perspectives for scientific argument studies

Now that scientific arguments can well be studied as contingent processes that emerge in conversations between scientists, four perspectives were born, depending on the context in which the argument is built. Keith and Rehg (2008) note that arguments are first constructed in local contexts where researchers talk and reflect privately. They refer to it as Local construction of arguments at the research site. Second, they move on to larger discourse communities, as a lot of the argumentation takes place in print. It is a context where scientific debate occurs as well. The authors call it Writing and controversy: science as discourse community and field. Third, as scientific ethos, funding mechanisms, disciplinary divisions, and other institutional and cultural components of science were noticed to have an additional impact on scientific arguments, the researchers encapsulated them as The Institutional Structuring of Scientific Argumentation. Fourth, the scientific argument-building process reaches the intersections of science and society, labeled as Public discourse and policy argumentation. In this paper, we only discuss the first and third perspectives because they both intersect with our case study. The first perspective and our study both involve the interest to examine arguments in situ and in the making, which in our case is the cancer research lab. The third perspective is relevant to our study as it focuses on how institutions, such as the cancer research lab, shape arguments. This aligns with our research findings that show how our scientists justify their claims based on institutional constraints, including practicing the status quo, adhering to cancer research standards, and being bound by the demand and supply interplay. However, the second and fourth perspectives are irrelevant to our study. The second perspective pertains to complete and explicit scientific arguments as presented to the scientific community, while the fourth perspective involves fully packaged arguments presented to the general public.

Local construction of arguments at the Research Site

Sociologists, anthropologists, historians of science, and communication scholars have made significant contributions to understanding local scientific practices, although cross-fertilization with argumentation studies is still minimal (Keith and Reigh, 2008). Latour and Woolgar (1979) look at how scientists create facts from data by converting qualifying assertions like “Smith noticed evidence for x” into unqualified factual claims like “x exists”. They describe scientists’ actions as driven by credibility rather than adherence to methodological principles. The rhetorical aspects of fact construction are further developed by Latour (1987). Comparing network parts with rhetorical resources, he explains the process of turning views into facts, where a “fact” becomes a proposition that no one has the resources to refute with a convincing counterargument. He connects lab-level argumentation with institutional and technological aspects of research. The present research, however similar to Latour and Woolgar’s (1979) in terms of intents, does not start from a verbalized argumentation. Rather, it provokes argumentation based on the observed tacit knowledge. Ethnomethodologists’ rigorously descriptive and detailed investigations of argumentation in the local context also help reveal the situated, local everyday logic (Lynch, 1993). Studies range from how data is interpreted in a neuroscience lab (Lynch, 1985), how mathematicians’ embodied social practices underpin claims on universality (Livingston, 1999), to how necessary an assertive style of communication is for an effective debate in high-energy physics (Traweek, 1988). Individual needs and goals, professional and other social interests, class, etc. appear to influence local argument-building (see MacKenzie and Barnes, 1979; MacKenzie, 1978).

The institutional structuring of scientific argumentation

Argumentation in science happens within particular institutional contexts under the influence of funding mechanisms, organizations, communication channels, modes of recognition, gatekeeping, etc. What uncovers these influences are interdisciplinary approaches and boundary concepts like ethos, consensus, logical discussion, and disciplinary boundaries. Most works on institutional structuring of scientific argumentation react to Merton’s (1973) “ethos” of science; norms such as universalism, organized skepticism, the pursuit of knowledge in an unbiased way, and communism guide scientists’ behavior and their collective effort. A significant problem that arises is the relationship between ethos and consensus. Gilbert and Mulkay (1984) see consensus as something discursively created within a context. Prelli (1989) argues that these norms are used as argumentative resources for establishing and undermining credibility. Hull (1988) demonstrates how institutional mechanisms promote collaboration among self-interested scientists. Paul Edwards (1996) shows institutional and cultural influences on arguments themselves. Taylor (1996) uses rhetorical analysis to demonstrate that scientific argumentation, is part of an “ecosystem” of people, journals, and organizations that certify or reject claims. The present research resonates with the current perspective, and more particularly with Prelli’s take. As we outline in our findings, the scientists (our informants) explain their practices by means of ‘standards’, which we find to be equivalent to Prelli’s ‘norms’. In their evidence, our scientists back up their claims, playing up to international standards in the field of cancer research.

Case study

Methodology

We used Potter and Wetherell (1987)’s Discursive Psychology as a linguistic and ethnographic theory and method. It focuses on discourse in contexts of use. The method comprises the following steps: Research question, data collection, observation and interviews, transcription, analysis, discussions, and conclusion. In Potter and Wetherell (1987), the authors stated that Discursive Psychology as a theory builds on assumptions from speech acts, ethnomethodology, and semiology. Using this method, our analysis does not see the role of language as to simply describe things, but also as to do them. This is what Potter and Wetherell (1987) call ‘doing versus stating’ role of language. More particularly, building on Austin (1962), this theory argues that a speaker with any utterance does simultaneously 3 sorts of things: (a) provides a specific meaning; (b) exerts a particular force; and (c) effects the consequences. Thus, our own analysis of the informants’ (cancer researchers scientists’) discourse considers the meaning, the force, and the effects they bring about within their own scientific knowledge.

Our inductive research question was as large as follows: what can be considered as an unstated scientific argument in this cancer research lab? Starting from the definition that we provided in the introduction section, we focus on the scientific practices that make assumptions without verbalizing them (i.e. tacit knowledge). Our corollary question is: how do we decide that this or that specific scientific practice is less than a full argument (not stated verbally or in writing) and more than tacit knowledge (because it comprises a claim and evidence, but shared nonverbally).

In terms of data collection, this ethnographic study was conducted in an oncology lab in Brussels. The lab includes 10 members and is part of a research pole made up of 4 labs. The lab members are postdoctoral and doctoral researchers, in addition to technicians and the director or chief investigator. They come from Belgium, France, Spain, the Netherlands, and Tunisia, and they hold university degrees from these countries. The lab director had served as a Post-doctoral fellow and research associate at the Brigham and Women’s Hospital and instructor in medicine at Harvard Medical School, Boston. As indicated on the lab website, they have made 255 publications and a number of patents, and they have achieved an H index of 81 in 2022.

Typically, the lab members work on cellsFootnote 1 —usually cancerousFootnote 2 —that they purchase from external providers. The different researchers in this lab usually aim at either understanding a particular aspect of the cell metabolismFootnote 3, or testing drugs on cancer cells to check their therapeuticFootnote 4 efficiencies. Any cells are immediately stored in the freezer as soon as they are delivered by the provider. If they left the freezer, the cells would soon die. When a researcher would want to start working on some cells, he or she would need to prepare for what they call cell cultureFootnote 5. Culturing cells means providing them with adequate conditions to get them to grow as they would in the human body. The cell culture starts with the preparation of the mediumFootnote 6. The medium is a liquid made of a number of nutrient ingredients for cells to consume that are supposed to substitute for what the human body consumes in terms of food and drink.

The cells, which are covered by the medium, soon get glued to the wall of the Petri dish taking the shape of a very thin whitish to colorless layer, which we can see with the naked eye. They should very quickly be placed in the incubator. The researchers change the medium from time to time, usually once every 3 days, because they say that the cells consume from the medium which becomes impoverished from a number of ingredients that the cells need, and also because the secretions let out by the cells need to be evacuated in order for the cells to consume cleaner medium. To do so, they suck the old medium using pipettesFootnote 7 and immediately inject the new one. These operations and all others are performed inside the hood (an apparatus used to hatch eggs or grow microorganisms under controlled conditions (Oxford dictionary)).

With regard to observation and interviews, the first author of this study conducted an 18-month ethnographic observation using Discursive Psychology (Potter and Wetherell, 1987). Based on the observation, we designed a semi-structured interview over the issue of medium usage, conducted with each participant individually, by the first author. Six months later, a collective interview, which went through the same questions, was conducted by the first author with the following 10 lab members. While the individual interviews were analyzed synchronically, the collective interview was analyzed diachronically. Therefore, our research had to be conducted inductively, for at least one obvious reason: aiming at studying unstated arguments, entails the nonexistence of these arguments prior to our observation. It is through observation and analysis that such arguments (expressed nonverbally) can be made explicit.

To protect the privacy of our participants, their names have been changed into the following pseudonyms:

  • Paul: The lab director/Principal Investigator.

  • Stephan, Daniel, Brian and Dilara: Post-doctoral researchers

  • Cédric and Gaelle: PhD researchers

  • Dalila and Marie: Technicians

  • Joane: a master student

The interviews were described using the Jefferson system of transcription notation, detailed in Attkinson and Heritage (1984). An outline of the system of transcription notation is included in the endnotes sectionFootnote 8. The next sections of this paper comprise the remaining elements of our method: analysis, discussion, and conclusions.

Identification of an unstated argument

The following unstated argument was identified through observation and explored through the interviews:

Claim: The medium usage is unquestionable.

Initial evidence: The medium usage follows the status quo, it adheres to standards, and it is the only possible way.

The medium is a liquid composed of a number of substances injected into the culture cells, on which the researchers do experiments. During the observation and also in the tape-recorded interviews, all the researchers told me that depending on the experiment, they are supposed to reconsider the medium components, knowing that some components should always be there (e.g., glucose). The first time I asked Brian about it, he told me that the medium is the nutrient for the cells and that “it replaces the food and drinks that humans consume”. The medium is usually pinky, especially when it is fresh (i.e., not yet injected into cells). After spending a couple of days on cells, the medium changes color and becomes less pink and more or less yellowish. When I asked why the color was changing, Stephan said that it was due to the consumption of different nutrients by the cells. He said, “If it is less pink, then it is more acidic”. Then he said that the medium has to be replaced regularly in order for the cells to be in a healthy and non-toxic environment. One technician from the cardiology lab told me that the change in color is also due to the fact that the cells discharge their waste and gases inside the medium, so the medium is no longer as clean as before. I understood that the medium is used constantly in any single experiment. Cells are never without a medium. I asked Sophie (a physician from the cardiology lab) if the medium was used everywhere else, in other labs. She said definitely because, without a medium, cells would die immediately.

I initially had two main concerns with this way of exposing cells to the medium. First, I wondered why the culture cells are constantly covered by the medium, unlike the cells in the human body. After all, humans eat and drink from time to time and are not totally and continuously exposed to nutrients. When I asked this question to Stephan, Brian, and Joanne, they did not seem to take it as a serious issue. In different ways, they all more or less said that the blood in the human body is actually feeding the cells in a constant way and that I should not compare the medium to the food and drinks but to the bloodstream. Bottom line this medium issue did not seem to have drawn the attention of the researchers. However, I had clearly contradictory accounts, which were reproduced during the interviews and which I will discuss in the next section.

My second concern was about whether the scientists knew the components of the medium. I started to inquire about it when I noticed that the usage of the medium is done in a mechanical way (i.e., unquestioned tacit knowledge). It is perhaps the most frequent operation that one can see in the lab. Many times a day, I saw the researchers or technicians grabbing bottles of medium from the stock and injecting a quantity onto the cells (which are placed in Petri dishesFootnote 9) using pipettes after removing the old medium. The suction of the old medium and the injection of the fresh one is operated in the culture room, inside the hoodFootnote 10. This operation, like many others, is done with great caution, because there is a risk of contaminating the cells with bacteria, fungi, or viruses. Strict rules are posted on the hood in order to remind the users of how to avoid contamination. Basically, everything that comes in contact with the cells or near them, particularly inside the hood, has to be sterilizedFootnote 11 immediately before use or has to be brand new (and which is supposed to be sterilized by the provider). The medium is the substance that is constantly in contact with the cells, which also means that it has to be permanently sterile.

Analysis: Evidence variation over the same claim

Through the following analysis, we wish to emphasize the scientists’ agreements on disagreements. In other words, they agree on their practices but disagree in terms of rationalizations.

Why are the culture cells constantly covered by the medium, unlike the cells in the human body?

Two positions

With regard to the fact that the medium is constantly on the cells, at least two clearly diverging accounts emerged right from the first question of the individual interview. A group of three participants made up of two postdocs (Stephan and Daniel) and the lab director (Paul) said that the system does not really mimic the human body, but for technical reasons, there are no better alternatives at the time being. Although the three of them mentioned that a new system has recently become known, the issue for this group is rather the unavailability of better technological or financial alternatives. For Stephan, it is a matter of practicality. Paul was very short and mentioned that it is “because we (.) cannot easily do otherwise”. When I mentioned the new system to him, he took six seconds of silence and mentioned its necessity, but also mentioned that it is still not well established. Daniel said in French almost the same thing as Paul, i.e., that the need is there, but the system is still not standardized.

On the other side stands a group of six participants (two postdocs: Brian and Dilara; two PhD students: Cédric and Gaelle; one technician: Dalila; and one master student: Joanne). This group has accounted with varying degrees for the fact that the current usage of the medium (being constantly on the cells) functions properly, mimicking how the human body consumes nutrients. With a slightly nuanced account, Brian said that “the medium mimics a bit the: blood (.) that feeds all the cells in the body (.) but in the blood you have like a range of glucose and you have less of this glucose you will start to take fat and usually start to generate glucose or energy (.) so there is an homeostasis in the blood and in the body and I think the medium tries to mimic this uh system (.) I I I wouldn’t say that in the body there are a lot of (.) very drastic change in the: in the food and in the concentration of nutrients”. With a rather more determined manner, Dilara said in French that this medium is doing a similar job as how it works in the human body.

Cédric was as determined as Dilara: “but your b:ody is quite the same (.) I mean you eat from time to time (.) so like you eat and then you take water but in your blood stream you are constant (.) nutrients level like you have in the cells like medium”. But Dalila’s account was at the end of the continuum. For her, the current system is not only close to how it works in the human body but there is no need for it to be that similar. Perhaps Dalila’s job as a technician requires her to master what is in force rather than think about alternatives not yet made available to her.

In a less determined way than anyone among the six group members who are at the beginning of the interview content with the current system, Gaelle said that the culture cells are constantly exposed to the medium, just ase the human body’s cells are exposed to nutrients all the time. For her, the body’s cells know how to get nutrients from the “stock”, but since the culture cells are not able to do that, exposing them to the medium in a constant way is a necessity. The most noticeable account among all of the participants is Joanne’s. Joanne belongs to this group of the six participants (those who first seemed content with the system), but she is in complete contradiction with them when she explains that the culture cells consume nutrients from the medium only when they need them, just like the body’s cells consume only when they need them. Basically, Joanne agrees with the second group but explains things according to the first group.

Positions reshuffled

Now, except for Dalila, the two groups (the one that seemed to be content with the usage of the medium and the one that said that they needed a better system) will collapse into one single group, as they will converge on the fact that this issue is of considerable importance. Stephan said that: “↑ I think it’s important (.) I think you have to (.) < if you are (.) every () of profiling enough clean medium uhm nutrients at regular intervals is important otherwise your culture cells are affected and >”. Paul said: “definitely (.) and I I I keep (.) every time I’m part of a jury (.) not this only in my field I keep asking the question are you sure that u:h u:h what you see is not influenced by the deprivation of your medium u:h in one nutrients?”. Daniel was as determined as Paul and said yes, this question is constantly on “their minds”; and that in their projects, they all the time think about what they put in the medium, in terms of quantity, quality, and dynamism.

From the second group, Brian said that the question is important because “it’s important you have to the: the same levels of nutrients more or less all the time”. Dilara said that the issue is important because we have to be sure that changing the medium will not skew the experiment. Gaelle provided the same argument on which the first group built their position (of not being satisfied with the system). Perhaps, Gaelle is figuring out the idea little by little: “yes (.) important not important but it’s uh yes it’s important as well […] no for me it’s more the::: the concentration of each stock not really the fact that it’s exposed all the time or not”. Cédric too seemed to be figuring out the importance of this issue at the time of explaining it to me, especially that he said in question two that he never thought about this issue. Cédric said “yeah it might be like when I was think for endothelial u:h cell there are in your life exposed to flow because they are inside the blood vessel and when you are cultivating cultivating it here in the lab they don’t have the same flow so may be important”. But Joanne was even more skeptical and withdrew her answer. She first said: “Yes”, but when I asked why, she said “hum::: (.) ↓ we because (.) it is important to know if we can u::h like ↓ I don’t know exactly how to say that (0.5) I don’t know (.) because if we: remove the medium the cell will die (0.4) I will have to think about it ((laughter))”. Finally, Dalila was again at the end of the continuum and was the only one who said that this question was not important. She said “no”, “because as long as we know that in cell culture we don’t totally mimic the reality, there should be no problem”. She also said that they do not ask such a question and that she personally never did.

Do the scientists, in reality, know the composition of the medium?

As I saw the participants using the medium in a mechanical way, I asked myself the question: do the researchers think about the medium, or is it one of the unnoticed elements of their research. It was not possible to answer the question just by observing or without discussing it with them, because they may think about that individually or during the design of the different research protocols. Therefore, I asked Brian, who told me that they do think about it very carefully and that their lab is specialized in the study of cancer metabolism, which means that these conditions are of particular importance to them. I addressed the same question to Stephan and then to Daniel who said the same thing. I then asked Joanne, who told me that the medium is usually purchased ready-made, but it can sometimes be made within the lab, using different raw ingredients that are themselves purchased. I went back to Daniel and told him that I wanted to further discuss this issue of the medium use. The discussion with Daniel made me conclude that unless the research is testing a particular ingredient of the medium, researchers usually use a standard one; a ready-made one that is available in the market. I asked Daniel and then all the lab members (except Dilara and Paul) if they knew all the ingredients of the medium that they were using by heart. They all said “NO”.

What came as a surprise to me was that the lab members do not know the whole list of ingredients in the medium in which the cells live, although they keep saying that every molecule can make drastic changes if we are not aware of its presence. Brian said (and he repeated this in the interview) that the list of the medium ingredients is too long to remember, but before deciding what medium to use, they think about it thoroughly. Further discussions with Stephan left me with three ideas: (a) most of the time the medium is taken for granted unless the particular research is specifically testing one ingredient; (b) the medium is only thought of, when they notice a problem, such as a strange behavior of the cells; and interestingly, (c) the participants are making efforts to show me that this is not an issue and that everything is under control. Bottom line, I had a feeling that the participants did seem comfortable discussing the medium with me. I decided to design a semi-structured interview dedicated to the medium issue because I suspected it to be an instance of trivialization—a trivialization of something that is otherwise described by the researchers to be extremely important. I was intrigued by the two extraordinarily paradoxical realities: (1) statements that any non-identified molecule in the medium would have implications on research and (2) not knowing what the medium contains.

Counter-argument emergence

Counter-argument

The medium usage is problematic.

Evidence of counter-argument

The medium usage is not optimally mimicked in vivo, it can systematically contaminate cells, and it can contain molecules that we ignore, which would result in systematic false negatives.

Our discussion over the initial two questions gave birth to a new inquiry about the possibility of the medium itself being a systematic cause of contamination, i.e., a virus in the medium contaminating cells. My question presupposed that the drug that cancer researchers test on cells is supposed to kill the disease and not the whole cell, but Brian and Dalila corrected me on one important matter. They said that most of the time the drugs are actually designed to kill a number of cells (some of them will anyway stay alive and keep dividing), and sometimes the drug is able to target the cancer cells specifically (leaving behind the non-cancerous cells). The aim is to keep killing the cells, not the disease (randomly or in a targeting way) until you reach a level where you do not see them proliferating anymore. Beyond this confusing idea, I will now address the remaining part of the question, related to the possibility of contamination through the medium. Stephan and, to a lesser degree, Brian and Gaelle, said that the medium as it is currently used can cause repetitive contamination, and therefore, if an experiment is testing a drug on the cells, it can be concluded that the drug is not effective, when it may well be, in fact, effective. The experiment would, in that case, be reported negative, when the problem is not in the tested drug, but in the medium that is skewing the experiment. The problem may then lie in the fact that the medium is contaminating the cells again and again. However, all the other participants do not see this idea of medium contamination as realistic.

To start with, Stephan initially confirmed the possibility of contamination through the medium, which is already half of the answer. I then clarified that I did not mean the medium exposed to cancer cells contaminating non-cancer cells. I clarified that what I meant was that the purchased medium contaminated cells again and again, while the drug is being tested. He said: “oh yeah for sure (.) it’s u:h there’s compounds that are excreted and are: that are floating in the medium and it’s also known that if you grow cancer cells or other cells you just only take the medium without the cells and transfer it to other healthy cells you might affect their function (.) that happens yeah that’s a known hormones grow factors”. Stephan also added that though such contamination is possible, researchers are able to know if the drug is not totally effective as a consequence of contamination.

Brian also agreed with the idea that contamination through the medium is possible. Basically, Brian also said that contamination is possible, but it takes time, and because the medium is changed every 2 days or so, the contamination may not have enough time to occur. When I mentioned that contamination in reality can happen before we can actually see it (two months, as he said), he admitted saying: “ah of course it starts before yeah of course (.) even after 24 h you can see changes”. Gaelle too said that this is possible, but then said “but we change the medium also”. So for Gaelle, if we change the medium, there is no possibility of contamination, although the medium is changed after 2 or 3 days, and according to Brian, contamination can occur in 24 h.

On the other side stand Daniel, Dilara, Cédric, Paul, Dalila, and Joanne. All of them do not see contamination through the medium as a possibility. Daniel said that this idea of contamination through the medium is definitely not a possibility. He said that contamination can occur through some residue from infected cells, but not through a molecule in the medium. He added that some researchers suggest that stem cells may be behind this contamination, but even this theory of stem cells is still not very well established. Daniel was confident that the medium could not be behind the contamination and the resistance. He said “no”. I said it has nothing to do with resistance. He said “no” again. He said “no” again.

Dilara too excluded the idea that the medium could cause contamination or that contamination is behind the cancer cells' resistance to drugs because the medium is changed every 2 or 3 days. This means that for Dilara, the medium contamination does not happen just because it does not have enough time to happen. She then explained that resistance is due to the fact that cells develop new mechanisms of adaptation against the drugs. However, when I asked her, if this adaptation is something that cancer research understands well (how it happens), she said that she herself does not know, but that it is becoming more understandable. Basically, Dilara has shifted the discussion from the possibility of contamination and resistance being caused by the medium into the discussion about the adaptation, which she said that she does not understand. Additionally, I said, “so, for you the medium cannot be in any case behind cancer cells resistance?”; She said “no, I wouldn’t say in any case, because the medium components may strengthen cancer” and then she developed her idea giving the example of the glucose to be possibly one of the factors of the cancer cells resistance. Bottom line, Dilara finished her explanation with the idea that the medium can cause resistance, but due to its composition, not through the persistent molecules.

Cédric said that he understands me very well, but that contamination through the medium is not possible because the medium does not contain the cancer molecule. Paul also said that contamination cannot be a matter of a molecule that stays in the medium or anywhere else. The cancer cells are already affected by the disease and have changed phenotype, which makes them proliferate and divide in an uncontrolled way without the presence of the cause anymore. Again, for Paul, cancer is a disease, a disorder, or a condition. He ruled out the possibility that this disorder is caused by a molecule that is always there. He is as confident as Cédric that the compound (or the molecule) is not there anymore.

Surprisingly, Dalila said that contamination is not possible, not because she does not think that the medium contains molecules, but because if the drug is designed to kill cancer, it would kill it in the cells and in the medium as well. Dalila was the first one to mention this idea, which now seems obvious to me. Then, Dalila stepped back from the idea that the molecule causing cancer could be in the medium. She said, “but for me cancer is not a molecule; cancer is the effect of that molecule on the cells,” and then she went back to saying that cancer is a molecule causing the proliferation of cells. Joanne’s accounts were similar to Dalila’s. While she agreed that cancer can be caused by a molecule, she did not accept the idea that the molecule would stay in the medium and keep contaminating the cells. She insisted that cancer resistance is not a matter of contamination but of adaptation. Joanne failed to explain what she means by adaptation, or how it happens, and neither had she demonstrated how we can know for sure that the molecules that cause cancer are no longer in the cells or in the medium. After rejecting my idea, Joanne stepped back and said that she does not really know: “No I don’t know may be (.) you can’t say for anything you don’t really know”.

There were two points that no one except Stephan was able to explain: (1) While I understand that cancer is not a molecule but a disorder that can be caused by many factors, including a molecule or a virus, how do we know that the molecule or virus that caused cancer has left the cells and is no longer there? It was very hard for me to try to solicit an account for this very specific point. No one of the participants was able to demonstrate that the molecule is no longer in the cells. Conversely, Stephan’s answer is not problematic to understand because he said that molecules that cause cancer may stay in the cells, may float in the medium, may cause auto-contamination, and may be behind drug resistance. (2) The second point is about the other causes of cancer resistance. All the other participants said that cancer resistance is due to adaptation, but they all said that they do not understand how adaptation works or what mechanisms cancer cells develop to be resistant to drugs. We will address this issue in the next subsection, within a detailed discussion about how each of the lab members built his or her accounts throughout the interview session.

Discussion

The mediumFootnote 12 usage in cancer-research cell cultureFootnote 13 encompasses an argument that has never been stated; an unstated argument that can hardly be grasped without ethnography. The unstated argument was implied and embedded in tacit knowledge type of practices (i.e. how scientists should use media in cancer research labs). While the scientists first described the medium usage as appropriate and non-problematic (the claim) through observation and in the interviews, they ended up considering that it can be serious (the counterclaim) and providing counterevidence. We will now summarize the evidence (presented to back up the claim) and contrast it with the counterevidence (which happens to contradict the claim):

Medium stagnancy

The scientists said that there are no better technical solutions to get the medium to flow and make in vitro cells consume nutrients in a way that is more similar to how in vivo cells do. Furthermore, when they were reminded of the existence of another system allowing the medium to flow (which was initially mentioned by one scientist during the observation), they described it as being closer to in vivo and hence better; but when they were asked why they did not consider using it, they minimized its usefulness. One participant said that it is not that important to be very close to in vivo, although in other contexts they always emphasized that there is no priority over being as close as possible to in vivo (i.e., to avoid false negative or false positive findings). Another scientist said other labs might be interested in the flow system, but not this lab in particular, because this lab has other priorities, while all her colleagues emphasized that this lab is, in particular, focusing on cancer metabolism and that the medium composition role is a core issue in this regard. On other occasions, the participants said that maybe it is better to use the flow technique and be closer to in vivo. We, therefore, deduce that the scientists fail to back up their initial claim with any evidence. They instead use personal opinions, whether or not they unanimously agree on details. Arguably, a personal opinion is not what we ideally expect from a scientist (although it may be presented side by side with expert-opinion), as it represents “a personal belief that may or may not be based in fact; a view or judgment that individuals form about something” (Frey et al., 2015, p. 11). However, the provided counterevidence showing that medium stagnancy can cause false negative or false positive results is an expert opinion, at least in Hitchcock’s (2005) terms. More precisely, the counterevidence comprises a warrant, as described by Hitchcock (2005) using Toulmin’s model: the gap between the in vitro and in vivo models is known to cause false negative and false positive results.

Auto-contamination

Only one scientist immediately recognized the risk of auto-contamination due to the current medium usage, providing counterevidence from the beginning. All the others initially refuted the idea that auto-contamination is possible, justifying their position by saying that cancer is not or cannot be caused by molecules (i.e., that it is only a matter of cell proliferation disorder). Thus, the initial claim is the following: auto-contamination emanating from the medium usage is not possible to happen, simply because cancer is not a matter of morbid molecules affecting the cells (evidence). Further discussions with some scientists made them say that while cancer is a disorder of cell proliferation, it can well be caused by morbid molecules and that these molecules can possibly stay in the medium and cause auto-contamination (counterevidence). One scientist never changed his mind during the individual interview but dramatically shifted his position during the collective interview. When it was pointed out that only one of them immediately acknowledged the auto-contamination risk and that others did not, the controversy emerged again, and this scientist managed to convince all his colleagues that the auto-contamination is quite possible and that cancer can be a matter of morbid molecules that never leave the cells and not only a state of proliferation disorder. After the “dispute” was settled, the scientists ended up acknowledging this auto-contamination risk; and the counterevidence managed to defeat the evidence and contradict the claim. Thus, the claim “auto-contamination is not possible” turns out to be based on a personal opinion, at least in Frey et al. (2015)’s sense outlined in the previous paragraph. However, the counterevidence that “auto-contamination can happen” is an expert-opinion because—after the provoked discussion—all the experts agreed that it was a possibility. Without the provocation of the discussion over the possibility of contamination through morbid molecules, the controversy would not be apparent and the dispute would not be settled. Such a settlement, which was the outcome of our provoked discussion, is what Walton (1998) insists on to make a rational argument: “the acceptance” as opposed to “belief” or “knowledge”.

Not knowing the medium components

The participants do not know all the medium components and they justified it differently: (a) that the list is long; (b) that they do think about them when they design research, but they do not know them by heart; and (c) that the medium components are kept secret by the providers. More importantly, although they mentioned that every detail may skew research results, the participants said that they trust the medium’s providers and that they do not check the components after them, because they are: (a) professional (i.e., specialized), (b) competent (i.e., accountable) and (c) trustworthy. When the emphasis was put on the question, “How do you know that they are trustworthy?”, they referred to their professionalism rather than their trustworthiness. When the basis of trustworthiness was questioned, they insisted that the lab is not able to do what the providers could do. Finally, when they were asked to account for the possible outcomes, if (for any reason) the providers happen to be systematically altering the medium’s components, they said that it would then be very serious: (a) “there you are blinded”; (b) “results would be wrong”; “(c) that would change a lot a lot”; and (d) “you can have terrible consequences”. However, as soon as they describe these negative consequences, they restate again that this scenario would not happen. When they were asked why again, they said again that providers have no reason to cheat. We, therefore, deduce that the scientists use their opinions to back up their claim, but have strong counterevidence that contradicts the claim. Again, the evidence of the initial claim (that the external providers are professional, competent and trustworthy) is based on personal opinion (in Frey et al. (2015)’s sense outlined in the previous paragraph). Conversely, the counterevidence (if the providers happen to be systematically altering the medium’s components, it would have severe consequences on their research results) turns out to be an expert opinion in Hitchcock (1998)’s terms, as it is well substantiated (to use the scientists’ terms again: (a) “there you are blinded”; (b) “results would be wrong”; (c) that would change a lot”; and (d) “you can have terrible consequences”).

Conclusions

The case study was all about an unstated argument that scientists assume to be unquestionable, and on which they base the totality of their research. The case pertaining to the medium usage in cell culture within a cancer research lab in Brussels. First, the usage of the medium for cell culture in cancer research, which was initially described as appropriate (personal opinion but totally accepted in practice), was later accounted for as possibly skewing experiments (expert opinion, but not practiced). The skew emanates from the nutrients’ concentration variations and the resulting cells' adaptation to this variation, which could potentially shift in vitro from in vivo models. Secondly, auto-contamination emanating from the current medium usage, which was initially said by most participants not to be a possibility (personal opinion not backed up), was admitted to be quite possible by one scientist who defeated the initial claim of all the others in the collective interview (expert-opinion backed up with a warrant and managed to convince all the other scientists). Finally, the participants trust the medium providers based on a personal opinion (i.e., they are trustworthy because they are serious, competent, and because they have no interest in cheating). Conversely, the counterevidence provoked by our research is based on expert opinion. It suggests that if the purchased medium turns out to be always different from how it is described, it can potentially skew their experiments (to use the lab director’s words, “there we would be completely blinded”).

We, therefore, wish to emphasize the need to study unstated scientific arguments as opposed to fully fledged science. Additionally, we wish to emphasize ethnography as the most appropriate methodological approach that enables social scientists to capture these otherwise ungraspable types of arguments. Furthermore, discourse analysis methods such as those used in the present study (Potter and Wetherell, 1987) constitute an effective methodological tool to capture the unstated scientific arguments and to pinpoint a scientific argument based on a personal opinion, which may be intertwined with an evidenced claim. Clearly, this case (summarized in the previous paragraph) exposes the basis or rationale of the scientists’ claims and distinguishes them from what can be considered their area of expertise. Moreover, targeting tacit knowledge as such is a strategic entry point to be able to identify unstated arguments. Finally, we wish to note that scientists, as humans, may, in addition to their mistakes and flaws, have their own personal or collective hidden agendas. Social scientists would therefore need to equip themselves with critical analysis skills. Such skills would enable them to explore rather than check scientific data. They should not be contented with data as facts but examine how they are produced. In so doing, looking at the socio-psychological foundations of the scientists’ discourses, such as in Mnasri and Papakonstantinidis (2020 and 2023), is a key analytical tool to better explore science and knowledge building at large. Hence, we suggest that scholars of argumentation studies would need to put a spotlight on the unstated arguments and not only be content with the verbally stated (written or spoken) arguments. It is quite important to note that this research is highly transposable, as according to our participants, their medium usage is identical to any other biology lab in the world. In other words, the medium usage is part of accepted practices and guidelines in cancer research worldwide. However, we should note that the rationale of the medium usage provided by our scientists (informants) cannot be generalized in at least two ways: (1) they cannot be assumed to represent the opinions of other scientists from other cancer research labs; and more importantly, and (2) they cannot be considered to represent the disciplinary narrative.