How Scientists Think?

Tenzin Pasang. June 13, 2020 10:19 PM. 3590

By Kalsang Tharpa*, PhD



This article is a result of a suggestion from a friend to give a presentation on “Scientific Methodology” to Tibetan monastic science scholars. So, my target audience will mainly be the monastic science scholars who are already engaged in the study of Buddhist philosophy.


On many occasions I have come across people making statements like “the person who repairs your watch or mobile phone is an experienced scientist”. To this, I must say that he or she may not necessarily be a scientist, let alone an experienced one. On the other extreme, a person who can speak to some degree about basic science and resemble your pre-conceived notion of a 19th century scientist also may not be considered a scientist. So, then who is a scientist? Scientists are people who use what is called “the scientific method” to advance their knowledge of the world. With this broader definition, you may also consider yourself a scientist based on your day-to-day thinking and approach in doing things. The key question here is – what is “the scientific method”? With only about ten years in my bag as a practicing scientist, specifically in the field of chemistry, I can only attempt to answer this question and my answer may or may not represent the entire gamut of scientific fields ranging from the study of logic to physical sciences to social sciences. You may, however, take my thoughts as a basis on which to build your understanding as you read and study more exhaustive books on this subject. 


In order to lighten up this topic for the more casual reader, which otherwise may require deep thinking, I am going to use an example which is entirely imaginary but will nonetheless highlight the points I wish to convey. There are two friends by the names of Lobsang and Tashi who are staring at the smoke that is coming out from a house. Tashi asks, "what caused the smoke?’", to which Lobsang replies that the smoke must be due to a fire in that house. Lobsang makes a logical guess about the presence of fire but to confirm it he must go inside the house to see for himself.


We can say from the above situation that there are three parts to the problem of ascertaining the presence of fire:

  1.      Evidence  - smoke coming out from a house.
  2.      Logical guess - there must be fire inside the house.
  3.      Experiment - going inside the house to check for fire.


The above forms the basis of scientific methodology. In short, one starts with the evidence, a logical guess is then made about the cause of the event, and experiments are conducted to determine whether to accept or reject the logical guess that was made. Although the basis of scientific methodology remains the same, its subtleness depends very much on the subtlety of the question you are trying to answer.


Let’s go ahead and consider a second question form Tashi – "what exactly is burning in that house"? At best, Lobsang can go and check for the object or material that is burning in the house but the situation may not be practically feasible to observe all the evidences. This also means that, in general, not all evidences can be proved by direct observation. Examples of this include the measurement of temperature of the sun, or mass of the earth, or distance between two galaxies, etc. So, in such a situation, in addition to the logical guess that Lobsang makes to answer the first question, he may also make some more assumptions. Making of such assumptions is known as hypothesis.


Let’s say Lobsang thinks that there is some relationship between the nature of the object and the resultant smoke, and he proposes a hypothesis that different objects on burning may give different colors of smoke or different density of smoke. To test his hypothesis, he may conduct some controlled experiments by burning different objects possibly present in that house. Suppose Lobsang observes from his experiments that burning wood gives dark thick smoke, paper light smoke and electrical appliances greenish light smoke. He will then have a certain level of confidence, if not a fully convincing one, to answer Tashi’s second question.


In order to understand how scientists do experiments to test hypotheses, I would like to take the example of tea making. Let’s say we step into the kitchen to make a delicious cup of tea. Whether a tea is delicious or not is subjective but an individual can compare today’s tea versus yesterday’s which depends exclusively on the procedure you followed in making the tea provided you used the same tea leaf and milk. Let’s say your assumed recipe for a good cup of tea involves adding one table spoon of tea leaf to a cup of milk and boiling for five minutes before adding the rest of the ingredients. This assumption can be tested by first varying the quantity of tea leaf (e.g., 0.5, 1, 1.5 and 2 spoons of tea leaf) and boiling for a fixed time of five minutes. Suppose you learn that 1.5 table spoons of tea leaf gives better taste. You would then fix the amount of tea leaf to 1.5 spoons while you study the effect of boiling time (e.g., 3, 5 and 7 mins). This type of experimentation involves varying one variable at a time. When there are many variables to change the number of experiments goes up as you try different combinations of the variables. This is where simulations, statistical analysis and combinatorial approaches allow scientists to change many variables at a time in their experiments.


From the above, Lobsang uses a subtler level of scientific methodology that comprises:

  1.      Refined observation - the nature of the smoke (color & density).
  2.      Hypothesis - different objects may produce different colors or densities of smoke.
  3.      Controlled Experimentation - burning different objects to look for different types of smoke.


There is yet another option for Lobsang where he can avoid performing the experiments to test his hypothesis. He can ask someone who has already done experiments on the relationship of the nature of smoke to the object being burned. He can seek this information from a large number of scientific journals where scientists periodically publish their findings. As we speak, there are about 30,000 scientific journals worldwide and roughly 5500 articles are being published each day! If you have internet, most of these journals are accessible through a keyword search and often requires payment to access the full texts. I must emphasize here that referring to just one scientific article may not provide a complete picture of a particular subject due to various reasons including insufficient data, overlooked or miss-interpreted inferences, or just because it was not in the scope of the study. Generally, a scientific article should lay out the assumptions made from past studies and the scope of the present study through sections such as “Title”, “Abstract”, “Introduction”, “Experimental Conditions”, “Results”, “Discussions”, “Conclusions” and “References”. As scientific knowledge is built upon multitude of other people’s findings, there is always a number of citations to past studies in any given article although there are exceptions. One example of this kind is a paper titled “On the Electrodynamcis of Moving Bodies” by Albert Einstein which was published in 1905 in the journal Annalen der Physik – this paper has zero citations. It is also rare to find pretty much exactly the same work being performed by different people and published separately. For example, if Isaac Newton publishes a paper on the law of gravitation based on the apple falling from the tree, you can’t publish the same law of gravitation just because you saw oranges also falling from the tree - they are essentially the same! Hence, scientific journals are primarily for publishing new findings and often improvements made over previous findings. They are global channels for exchanging ideas and critiquing each other’s findings while pushing the boundaries of human knowledge.       


In case, Lobsang does not find the articles he is looking for in scientific journals, he can write one based on his findings and submit for publication. The stages of publication involve peer review by experts, back-and-forth questionings on the validity of his experiments and finally, if found convincing, the paper gets published in the journal. Thereafter, it undergoes different levels of validation by other scientists, from reproducing his findings to applying it in real-life cases. Thus, the conclusions made by Lobsang may change or a more well-defined and specific conditions would be set around his conclusions. For example, it may look like – only if the paper is of a particular thickness and burnt within confined space of a certain dimension would it generate light smoke, while outside of these conditions it may generate some other types of smoke. Such critical analyses backed by experimental proofs are greatly appreciated and accepted in the scientific community. Hence, Lobsang’s work will undergo rigorous scrutiny within the community of scientists.


So, we introduce here yet another important step in scientific methodology, namely reading scientific journals. Hence, a better and a more intelligent way for Lobsang to answer Tashi’s second question would involve the following:


  1.      Refined observation - the nature of the smoke (color& density).
  2.      Hypothesis - different color or intensity of the smoke may point to different objects burning.
  3.      Literature survey - searching for answer in scientific journals.
  4.      Controlled Experimentation - burning different objects to look for different types of smoke.


The more subtle the question you are seeking to answer, the deeper scientific investigation it requires, and soon I will show its implications in advancing our knowledge. For now, let’s take the third question from Tashi – “why would different object give different colors of smoke when burned? In comparison to the first or the second question, the third question requires deeper analysis. In search of the answer, Lobsang may need to make more than one hypothesis and follow a slightly different approach from that of the previous two cases. At this point, you may observe some iterations between hypotheses and experimentations but the basic philosophy of scientific methodology remains unchanged.


Firstly, a process called hypothesis testing is conducted wherein a number of hypotheses are proposed and each of them are tested either through logic, prevailing laws, theories or possibly through quick experimentations. This process ranks the different hypotheses by their credibility and a number of them may get rejected at this stage.


Secondly, Lobsang may use deductive or inductive reasoning in his investigation. He is using deductive reasoning if he investigates why burning different objects generate different colors of smoke from the smoke’s perspective.  In contrast, he is using inductive reasoning if he investigates the same question by first trying to understand the nature of the object. He may use both approaches to validate or invalidate his hypotheses. For example, if Lobsang finds that the presence of chemical ‘A’ in the smoke results in its blue color then he ascertains that by analysing the presence of chemical ‘A’ in that very object. Before he makes a conclusion, he validates his findings by burning chemical ‘A’ in isolation.


As the process of investigation continues, answer to one question may raise another question and so forth. The emergence of a series of such questions in scientific investigations is what keeps the scientist excited and busy at work. At this stage, Lobsang could carry on with his investigation by asking the following questions – what is there in chemical ‘A’ that gives off blue color smoke and under what conditions? Suppose, chemical ‘A’ imparts blue color smoke under all normal conditions then Lobsang would go on to find the phenomena or the underlying nature of chemical ‘A’ responsible for blue color smoke. Generally, a chain of such questions would lead a scientist to understand the very nature of the relationship between the object and the color of the smoke, at least so in this example. Otherwise, in general, as Albert Einstein put it rightly, “the important thing is to never stop questioning”. A chain of such questions would reveal the nature of the phenomena - it is like peeling an onion, layer by layer.   


What would Lobsang do if he thoroughly understands the mechanism by which chemical ‘A’ gives blue color smoke? He would then test his mechanistic understanding on other chemicals. If his mechanistic understanding is correct under a given set of conditions and found answerable to why chemical ‘B’ give yellow color smoke whereas chemical ‘C’ gives colorless smoke then it becomes a theory. Hey may call it “A Theory of Colorful Smoke”! Using this theory, he can answer most, if not all, types of questions pertaining to what object is burning in that house simply by the color of the smoke emanating from the burning house. Most importantly, using his theoretical understanding, Lobsang can make predictions about what would be the color of the smoke for chemicals that he hasn’t studied yet. It is like how Einstein’s Theory of Relativity accurately predicted the bending of light due to the gravitational pull way before the experimental proof came from the scientific observation led by Sir Arthur Eddington.


If somebody makes a device with a camera that detects smoke and shows you in display the type of object being burned by using the theory proposed by Lobsang, then he is developing what is called technology. Technology is a result of the application of scientific theories or principles. Over a period, you may see different versions of this device with more sensitive camera or display with touch screen and all that, but the underlying principle remains the same for the device. Each time, technological advancement is taking place, not necessarily scientific breakthroughs. Therefore, going back to what I stated in the beginning of this article, the person who knows how to use this device or knows how to repair it may not necessarily be called a scientist.


In a nutshell, the stages of scientific discovery involve questioning, hypothesis testing, experimentation, theorizing and application (birth of a technology). We saw that hypothesis, experiments and theory undergo rigorous testing before being accepted by the larger scientific community but do we do the same kind of tests on the question we asked in the first place? I think the scientist does not reject any questions but classify them into answerable or non-answerable. Questions such as “did God create this universe or not” cannot be answered through the scientific methodology. That is because you cannot prove or disprove the existence of God. But at the same time, scientists do often modify the question to make it answerable through the scientific approach. He could ask instead - “how did the universe form?” Questions like this gave rise to some of the most important theories such as the Big Bang Theory and the Theory of Evolution.


Now a question you may ask is – “do scientists always rigidly follow the steps of scientific discovery starting from a question to the birth of a technology”? To that I would say - not necessarily. For instance, the drug aspirin, one of the most widely used drug in the world, was first sold in 1899 just because it was found to be effective in reducing body pain. But, as we continue to consume 50-120 billion pills per year, we still have many conflicting theories about how exactly aspirin works in our body. Likewise, there are processes developed in the laboratory that result in the desired product but scientists do not know the exact mechanisms of how that product is formed. The famous example is the making of urea, a synthetic fertilizer that led to food surplus in the world.  We have been making urea since 1922 and two Nobel prizes in chemistry have been awarded – first one in 1930 for developing this process and another in 2007, almost eight decades later, for mapping the detailed process of making urea!


It is also not true that scientists do not believe in something until they can see it. Not all experimental proofs can be seen under the microscope or telescope. Many of them are either inferred or detected indirectly. For example, a theoretical physicist may follow scientific methodology but the conclusions of his study may be based entirely on the accuracy of the mathematical models that he is using. When Albert Einstein and his wife Elsa were touring Mount Wilson Observatory in Southern California, Elsa happened to ask one of the astronomers there what those giant telescopes were used for. When the astronomer said those powerful telescopes were used to help reveal the secret of the universe, Elsa laughed and replied, “My husband does that on the back of old envelopes.” So, it is possible to not only mathematically prove or disprove a hypothesis but also build a theory based on carefully-built mathematical models.


While a work carried out using scientific methodology can reveal the true workings of physical phenomena, it is possible that you may encounter non-validated theories or unintended false findings in scientific journals. I wanted to give you an example of this before I conclude this article. Protonosil, an antibacterial drug, was first synthesized by two laboratory scientists from a company called Bayer. Protonosil is a modified dye compound and those two scientists knew from a series of experimentations that it is the modifier which has antibacterial property not necessarily the modified dye. The modified dye worked just because the modifier was present. However, the company decided to sell Protonosil- a modified dye, not the modifier, because they were in dye business. It was much later when a scientist from another company pointed this out and the sale of Protonosil plummeted as a result. I wanted to point out here that the advancement of scientific knowledge entails costly experimentations and their results may be influenced by the entities funding the study. In another scenario, you may come across accurate scientific results from a study but the studies may have a limited scope. Thalidomide, a drug used to cure morning sickness ofexpecting mother, was first marketed in 1957. By 1961, it was discovered that the drug had caused around 4000 cases of infant deaths and 6000 cases of birth defects. The scientists might have reported accurate and positive results on use of Thalidomide for curing morning sickness, but the fact that it caused other problems was unexpected and resulted from the limited scope of the initial studies on the drug.


In conclusion, I would like to emphasize that the explanation of scientific methodology in this article may have been influenced disproportionately by my association with physical science. There are broadly other two branches of science, namely logic/mathematical science and social science - experts in those domains may have similar or different interpretations about how scientists think and derive conclusions. Although there are communities of experts within a scientific domain that, by and large, work collaboratively, a large number of scientists work in very specialized, often lonely, fields. This might mean that the scenario of Tashi asking a series of questions to Lobsang, as in this article, would be in fact like Lobsang asking the questions to himself in his lonely scientific pursuit. Finally, to gain further understanding of the scientific method, you should read or talk to other scientists. There is, however, no better way to understand the scientific method than to study science and experience it yourself!




Special thanks to Dr. Tenzin Choephel la for editing and turning this article into a much more readable form. I would like to thank Gyen Karma Rinchen la, Dickey Wangmo la and Dr. Tenzin Pasang la for providing valuable feedback. I must also thank the two young monks (depicted as Lobsang and Tashi in this article) for giving me permission to use their photo which was taken by Ven. Ngawang Norbu la.



* Kalsang Tharpa is a researcher at Shell Technology Center in Bangalore, India.  


Leave a Comment

Most Popular Articles