What is Science?

Skeptical Medicine

John Byrne, M.D

What Should Science Do?

   ** Fact:
   ** Law:
   ** Theory:

The Scientific Method




Absence of Evidence is Not Necessarily Evidence of Absence

Balancing Confirmation and Falsification

Putting New Information in Perspective

Types of Science

Consensus and Paradigm Shifts


References and Links

What Should Science Do?

Science is a human endeavor for systematically accumulating reliable knowledge.

As discussed in the Philosophy and Science section, science is a method concerned with “is” questions. Science is also a collection of “is” statements (statements about how the world may actually work) obtained through systematic methodology. It allows us to create models of reality that we can use to make accurate, reliable and reproducible predictions. We continuously question, modify and improve these models as new information is discovered. **Science does not answer “should” questions. Philosophy answers “should” questions and is informed by science. As health care providers, we are interested in human affairs and human well-being. We ask ourselves “should” questions all the time. Should we prescribe a drug? Should we recommend surgery? Should we push for therapies whose goal is quality of life over quantity? We answer these questions by considering values, but also, by considering the knowledge provided by science. Science is a human endeavor.

Prof. Robert Hazen of George Mason University states:

the role of science is to formulate evermore exact descriptions of the physical world and its events. Some situations, like the orbit of a planet are amenable to exact mathematical descriptions. Other situations, like the flipping of a coin are better described in terms of probability.The real advantage of science is that it constrains us to describe the world as it IS, not as we wish it to be.

A fundamental assumption of science is that nature actually exists in a single, coherent way. The laws of nature apply everywhere. If an explanation for an observation seems to violate an established scientific principle, then Occam’s Razor would suggest that an alternative explanation should be sought or that our observation is in error. Employing these assumption has been more successful than the alternatives. Since we choose to value induction (see the Philosophy and Science section), we conclude that science is the best source of knowledge to inform our decisions whenever possible.

“Scientific knowledge is a body of statements of varying degrees of certainty — some most unsure, some nearly sure, none absolutely certain.”

Richard P. Feynman

Science is not certain. Scientific knowledge is different than the type of knowledge sought by philosophers of past ages. It can never represent ‘Truth’ in the metaphysical sense. Scientific knowledge represents models of the natural world that can be used to make predictions. As our knowledge grows, the predictive power of the models grow. All scientific knowledge is provisional. It may change as new information and understanding arises. Even ‘facts’ are provisional. This gives the scientist and the scientific skeptic power, for, unlike other types of knowledge, the scientist can at least understand his own uncertainty and use it wisely. No knowledge, outside of the awareness of one’s own existence and feelings, has certainty. To state that one is absolutely certain about an empirical claim is folly.

“The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.”

Bertrand Russell

Science begins with ideas, not conclusions. Conclusions only come after the rigorous processes of the scientific method have been completed properly. The conclusions of science are always tentative. Pseudoscience begins with preconceived conclusions, and attempts to defend and confirm them at all costs. The conclusions of pseudoscience are therefore not tentative. They are not science. They are usually wrong, or “not even wrong“. Preconceived ideas are fine. Preconceived conclusions are never part of science.


Stedman’s Medical Dictionary describes science as: The observation, identification, description, experimental investigation, and theoretical explanation of phenomena.Science allows us to accumulate knowledge. However, as we must constantly remind ourselves, scientific knowledge is provisional. It is not certain. There are several types of scientific knowledge. Fundamental misunderstandings of their definitions has needlessly fueled many debates.

The terms fact, law, hypothesis, and theory are used frequently and sometimes interchangeably in day-to-day conversations. But for scientists, they have very specific meanings. When speaking scientifically, these special terms define what we mean by scientific knowledge and evidence. The following definitions were developed by members of the National Academy of Sciences:

** Fact:

An observation that has been repeatedly confirmed.

In science, a ‘fact’ refers to a single thing that is observed to be the same way every time it is observed. Psychologist and historian Michael Shermer wrote, “A claim becomes factual when it is confirmed to such an extent it would be reasonable to offer temporary agreement.”

A scientific fact, such as the mass of a proton, is provisional. We accept it because it is always consistent. Our certainty about scientific facts are limited by our ability to observe them, and by the inductive process of inferring a general rule from independent observations (eg. – every time we measure the mass of a proton, we observe it to approximately 1.6726 x 10-27 kg; therefore this is its mass). Facts are the raw data of our scientific knowledge.

** Law:

A descriptive generalization about how some aspect of the natural world behaves under stated circumstances.

A scientific ‘law’ is a confirmed statement that quantifies and generalizes the behavior of natural things or processes. Paleontologist and author Steven Jay Gould described a scientific law as “a generalization so overwhelmingly confirmed by empiric observation that it would be perverse to withhold provisional assent.” Like facts, we arive at laws through an inductive process of inferring a general rule from independent observations (eg. – every time we measure an object’s force (F), it equals its mass (m) times its acceleration (a); therefore F=ma). Laws provide the formulas of our scientific knowledge.

** Theory:

A well-substantiated explanation of some aspect of the natural world that can incorporate facts, laws, inferences, and tested hypotheses.

In science, a ‘theory’ is the most important kind of knowledge. A theory explains “why” something “is”. It’s often the current product of decades (or centuries) of scientific research; research that incorporates numerous observations about the scientific phenomenon it is trying to explain. Germ Theory encompasses all of the facts and activities of microorganisms and the diseases that they cause. It explains how infectious disease happens. Such a theory leads to many testable statements that can verify or falsify aspects of the theory.

A scientific theory is understood to be a theory that has been so rigorously tested and supported by numerous lines of inquiry, that any single disconfirming observation will likely not negate it. Scientific theories are conceptual frameworks that act as explanatory models for some aspect of the natural world. Components of the framework may be challenged and found false, however this usually leads to a modification and an improvement in the theory in question. Theories evolve with new ideas and new evidence. Theories are rarely complete, but as science progresses, they may become robust.

Opponents of science, sometimes called “denialists”, often use phrases such as “It’s only a theory” in debates with science advocates. Such chicanery equivocates the scientific usage of the word “theory” with dismissive words like “hunch” or “guess”. As we can see, this is erroneous; a well established theory is quite different. A theory, especially a foundational theory, represents the best current explanation for a scientific phenomenon.

When a skeptic asks for evidence, he or she is really asking for examples of the above types of knowledge.

** Hypothesis:

A testable statement about the natural world that can be used to build more complex inferences and explanations.

This is listed here last because, unlike facts, laws and theories, hypotheses really are not knowledge as much as conjecture. As the word implies, they are hypothetical. Once thoroughly tested and tentatively accepted, a hypothesis may flourish into a new theory or become part of the supporting structure of an existing theory.

Most school children learn that a hypothesis is an “educated guess”. This is true, but it takes on a more relevant meaning when put into context with respect to the meanings of ‘fact’, ‘law’ and ‘theory’. In science, a hypothesis is generated inductively as an idea to possibly explain observations. A hypothesis may be thought of as a proposed idea that may fill in a gap in our knowledge.

Some would say that a hypothesis is an untested theory. However, this veers from the strict definition listed above. Hypotheses may flourish out of scientific theories as well as lead to them. For example, if “Germ Theory” is true, then maybe ‘Syndrome X’ is due to a virus.

Hypotheses should be at least potentially testable to qualify as scientific.

Hypothesis formation is inductive. We build them from as general statements from specific observations, facts, laws or theories. Hypothesis testing is deductive. It is a statement in which the conclusion follows necessarily if the premises are true.

Hypotheses that survive testing can network together to form theories and laws. Also, facts, laws and theories can form the premises for scientific hypotheses. Upon considering known facts, laws and theories, we induce new ideas that are testable. A hypothesis test is a statement of such an idea that can be expressed in an “if…then” form. For instance, “If Theory A is correct, then, under certain circumstances, we should see Data B. This forms the basis for experimentation and the scientific method.

The Scientific Method

Scientific knowledge is accumulated through a rigorous method. People have cognitive biases that cause them to favor some ideas over others. Because of these biases, we are often unable to objectively assess the relative truth of our ideas. Throughout most of recorded history, we did not have a reliable method for sorting these ideas by order of their relative truths. We needed a method for valuing one idea over another and for rejecting bad ideas altogether. Such a method would allow for the accumulation of knowledge in the form of facts, laws, and theories. We call this the “scientific method”.

The American Heritage Dictionary of Cultural Literacy defines the scientific method as: An orderly technique of investigation that is supposed to account for scientific progress. Its origins date back to the 17th century with Sir Francis Bacon. He described a sequence of steps to investigate potential causes (“forms”) of phenomenon. History knows these steps as the Baconian Method. In the 18th century, philosopher William Whewell expanded upon Bacon’s thoughts in his Hypothetico-deductive model.

Currently, the scientific method is generally recognized to proceed with the following classic steps (Note – these are paraphrased by the site author):

1) Careful observations of nature (gather data).

2) From this specific data, infer a general explanation or relationship for the data. (Induction, hypothesis)

3) Make new, specific predictions based on these theories and laws. (Deduction, hypothesis testing)

4) Perform proper experiments or observational tests to check the validity of the predictions.

5) Make logical conclusions from the results.

New ideas are generally formed through induction from observations or previously obtained knowledge. Richard Feynman once said that, in fact, it doesn’t really matter where the ideas actually come from. It is the method of testing those ideas that make science. Dr. Feynman was correct, however, one could argue that most original ideas occur through inductive reasoning.

One may consider the initial, inferred general explanation to be an ‘early’ or proposed theory. Such an idea must prove through proper experimentation that it can consistently make useful and reliable predictions before achieving the full title of “scientific theory”.

(Note – ideas that provide reliable relationships rather than explanations for the data would be labelled as “laws”).

Of course, these are general rules, and science does not always proceed in such an orderly fashion. Indeed, some subjects (like astronomy) are mainly observational and do not often lend themselves to controlled experiments. Knowledge obtained from these fields is no less scientific.

One may add that the nature of the process (observations, inferences, hypotheses, experimental methods, results and conclusions) must be defined, transparent and unambiguous for peers to review, scrutinize and reproduce. Science is a social process. Peers must be able to independently critique, reproduce and verify the results. Arguably, these additional steps are at least as important as the classic steps.

It may be helpful here to introduce an important axiom to distinguish scientific ideas from non-scientific ideas. This axiom is called “falsifiability”.


The philosopher, Karl Popper, distinguished scientific ideas from other ideas with the notion of Falsifiability.

**Falsifiability is the logical possibility that an assertion can be contradicted by an observation or the outcome of a physical experiment. That something is “falsifiable” does not mean it is false; rather that, if it is false, then some observation or experiment will yield reproducible results that conflict with it.

Essentially, an idea is scientific if it can potentially be tested and disproved. A scientific notion takes risks in that we can use it to make very specific predictions, which if not found, could potentially falsify it.

A falsifiable idea that is tested and retested, yet never found to be false, survives as a scientific idea. With this line of reasoning, a scientific idea can never be 100% proven. However, it can become more robust. If found to be false, then it must be modified or abandoned. A notion that has been falsified, yet still adhered to by its proponents, belongs to the world of “pseudoscience“. Proponents of pseudoscience may attempt to avoid the falsification label by making ad hoc modifications to their theory such that any disconfirming evidence can be explained away. Modifications of this sort render the theory untestable. In other words, the theory becomes unfalsifiable. Logicians call this “special pleading”. According to Popper, the idea becomes unscientific, and hence, cannot be labelled a scientific ‘theory’.

Some theories lead to predictions of very rare events or events that are difficult to observe. Attempts to falsify such theories may prove difficult. If one does not observe a very rare event, then we cannot really say that the idea is falsified. However, to be classified as scientific, the idea has to potentially be falsifiable. This is discussed more in the section below.

In science, it is acceptable to modify one’s theory to better explain the data, but the modification must allow for new, specific predictions to be made. The modified theory must also be potentially falsifiable. In this way, legitimate theories evolve and strengthen.

Ideas that are not at least potentially falsifiable are not science. They may be philosophic, such as in ethics, aesthetics, politics or religion. Science does not have much to say about such ideas. Debates in these areas lie in the philosophical domains of reason, logic, and argumentation. However, if a proponent of such ideas makes a claim that is potentially falsifiable, then it becomes fair for the skeptic to ask for evidence.


Some ideas are not falsifiable, yet they cohere with current scientific knowledge. Since they are not falsifiable, they cannot be used to make predictions, and therefore cannot be classified as scientific knowledge. Many refer to such ideas as “protoscience“. A common example in physics is String Theory. Currently ‘String Theory’ offers no implications that can be tested, however there are no components of ‘String Theory’ that conflict with current scientific knowledge. Technically, String Theory does not meet our scientific definition of the word “theory”.

In medical science, “Germ Theory” was likely once a protoscience, however, with further understanding, falsifiable claims were devised and tested. In the 19th century, physician Ignaz Semmelweis made observations about the condition ‘peuerperal fever’, which caused illness in new mothers on maternity wards. He used inductive reasoning to infer that the disease was caused by invisible agents on the doctor’s hands. He then deduced (hypothesised) that if providers washed their hands, the incidence of the disease would be reduced. He collected data and showed that this was indeed the case. Although it was not recognized in Semmelweis’s day, at that moment, ‘Germ Theory’ was promoted from protoscience to science. It is now a foundational theory in medicine.

One may argue that diagnoses such as fibromyalgia and chronic fatigue syndrome may fall under the heading of ‘protoscience’. They are defined by descriptions of symptoms, without a known biologic mechanism. As such, they are not really falsifiable. They are defined by their definitions. This circular reasoning has lead some to be skeptical. However, our understanding of these ideas may evolve (as with ‘Germ Theory’) and they may one day become true science.

We make this distinction between science and protoscience not to stifle protoscience, but to show that non-falsifiable ideas may one day turn out to be good ideas. At the moment though, we have no way of knowing. We further differentiate science and protoscience from “pseudoscience”, which represents ideas that have been falsified, or ideas that are not falsifiable, but deceptively promoted as useful knowledge nevertheless. Pseudoscience also includes non-falsifiable ideas that are incoherent with current scientific knowledge. (See the section on Pseudoscience).


From here, we could add some potential outcomes to the scientific method:

After one’s results are presented to relevant experts for review and possible replication…

… if the results repeatedly contradict the predictions, then we may say that the theory is “falsified” (or at least there is a problem).

… if the results agree, then the theory is supported.

… if no results are found, then the theory may be weakened, but not falsified.

… if the theory survives attempts at falsification, then it gains strength.

… if a theory is supported by a convergence of evidence from multiple lines of scientific inquiry, then it becomes foundational to our understanding of the world.

It should be noted that falsification of a foundational theory would require the falsification of numerous lines of supporting evidence. Some aspects of a foundational theory, like Germ Theory, may be wrong. In fact, it would be surprising if all aspects of a robust theory were correct. If disconfirming evidence is found against one line of support for the theory, then only that line of support would be falsified. Doing so forces the modification of the theory, and hence, makes it stronger. The burden of proof for overturning an entire, successful, foundational theory is an incredibly large burden indeed. But still, it is not impossible.

Absence of Evidence is Not Necessarily Evidence of Absence

One may form a hypothesis that predicts a commonality across a set. The statement, “All swans are white” is potentially falsified by finding one black swan among the set of “all swans”. Perhaps one would search the globe trying to find a black one, but never find one. Unless the searcher can account for the entire set of “all swans”, she may only ever succeed at confirming the theory. (See “The Black Swan Fallacy“)

One may form a hypothesis that predicts the existence of a rare event. When tested, one may only get equivocal results. A paleontologist may try to prove that a certain dinosaur lived in a certain valley. She may or may not find it by digging in random spots. If she works all of her professional life without success, the theory is neither confirmed or falsified. It may be weakened, but not exactly falsified. At that point, Occam’s Razor guides us to favor the “Null Hypothesis”.

** The Null Hypothesis is the negation of the hypothesis being tested. While a hypothesis may look to confirm that Theory A is true, the Null Hypothesis essentially states that Theory A is not true. If never confirmed,Theory A becomes very unlikely, and it may be wise to abandon it in favor of the Null Hypothesis. In medical treatment studies, the Null Hypothesis essentially is the position that the treatment in question has no effect. Such studies end with either the rejection or acceptance of the Null Hypothesis.

In science and skepticism, we are advised to take the position of the Null Hypothesis to be the default position. Most ideas turn out to be wrong. We start from the position of the Null Hypothesis, then test compare its ‘fit’ of the data with the ‘alternative hypothesis’. The ‘alternative hypothesis’ is our new idea. If our new, alternative hypothesis explains the data better than the Null Hypothesis, then we “reject the null hypothesis” and tentatively accept the new idea.

This axiom is the reason that theories cannot be 100% proven. If a potentially falsifiable theory has been repeatedly confirmed and has survived legitimate attempts to falsify it, then it becomes robust. But the potential for contradictory evidence remains. This, however, does not diminish the power of a well established theory.

Balancing Confirmation and Falsification

Experiments designed only to confirm a hypothesis or theory are hallmarks of pseudoscience.

For instance, if one thinks that the Earth is flat, one could design an experiment that only samples flat parts. Science would not advance if we only used a method designed to confirm what we want to be true and ignore the possibility of being wrong. It is natural to seek only confirmation, especially if we have an emotional attachment to an idea. This is called “Confirmation Bias”. A researcher may also inadvertently collect data only from subjects that would likely confirm a favored idea. This is called “Selection Bias. (See Cognitive Biases). **Confirmation Bias and Selection Bias drive pseudoscience.

Karl Popper’s approach to the scientific method embraced the idea of falsification, not only to differentiate science from pseudoscience, but as the goal of scientific methodology. Since theories can not be absolutely proven, he felt that scientists should focus on falsifying them. For instance, a phenomenon may be explained by theories A, B or C. We should try eliminating them through falsification. If contradictory evidence is found for B and C, then A becomes more robust. Science often proceeds this way.

However, as with the case of the ‘White Swan Theory’, hypotheses that predict an absolute commonality among members of a set may never actually be falsified. The contradictory example may never be found. Such theories may be considered to be strong, but in reality, they could be wrong. One has only to find the single black swan.

As with the case of the paleontologist above, hypotheses that predict rare events may never be confirmed. The object of the prediction may never be found. Such theories may be considered weak, but in reality may be correct. One has only to find the dinosaur bone.

Because scientists are human, researchers tend to produce studies that confirm their pet theories. Studies that confirm popular ideas may get published more frequently than negative, equivocal or disconfirming studies. However, because science is competitive, other scientists are often motivated to find disconfirming evidence against the popular theory. Originators of new theories tend to seek confirmation, but their competitors may tend to seek falsification.

Hence, science does not depend on any one study or one researcher, which in the short term, may muddy the water with conflicting reports. But the balance between confirmation and falsification is a group effort. Science is a group effort. The process makes uneven progress and is self-correcting.

Putting New Information in Perspective

New ideas are common. They fuel new research. Without new ideas, science would stagnate. Logic would have it that new ideas that cohere with established knowledge would likely be accepted with little scrutiny. Such new ideas are usually fairly small and rarely lead to paradigm shifts in science or fame and glory for the scientist. Plausible, small ideas are extremely important. They are the “bread and butter” of science. Laboratories across the world continuously plug away studying these small, plausible ideas. The sum total of their efforts strengthen the foundations of science and lead to slow-but-sure progress.

Logic would have it that if a new idea contradicts established knowledge, or depends on unestablished assumptions (an ‘extraordinary claim’), then the evidence for such a claim should be strong, unambiguous, reproducible and held to scrutiny (‘extraordinary evidence’).

Another important concept is “prior probability“. The likelihood of a theory being true may change in light of new evidence or competing theories. Prior probability refers to the theory’s probability of being true before considering new information. A strong or foundational theory is well established by multiple lines of evidence. It is accepted widely and hypotheses based on it are extremely successful at making accurate predictions. One would say that a foundational theory’s prior probability is very high. It would take a mountain of undeniable, disconfirming evidence to overturn it. On the other hand, a new theory that is incoherent with established laws and theories would be considered to have a very low prior probability. It would take a mountain of undeniable confirming evidence to support it. Some theories are somewhere in the middle. New data could very likely strengthen or weaken them.

Bayesian probability is often used to compute the relative probabilities of ideas with respect to their prior probabilities and new data. The strength of a theory can be increased or weakened in light of new evidence.

Thomas Bayes proposed a simple mathematical relationship for the probability of a thing in light of our prior understanding and new evidence. A Bayesian approach to rating scientific ideas is based on Baye’s Theorem.

Proponents argue that science, particularly medical science, should utilize the Bayesian approach as the preferred methodology for determining the relative value of theories precisely because it factors in prior probability with new data. As such, a small amount of confirming evidence for an seemingly implausible claim would not seem as significant as it would otherwise. Remember, extraordinary claims require extraordinary evidence.

Bayesian probability essentially states the following:

The New Probability of a Theory is proportional to its Prior Probability x the Strength of New Evidence.

Thus, the likelihood of a theory being true fluctuates in light of new evidence. Strong theories have strong foundations and high prior probabilities. To overturn such a theory would require new evidence that is ‘extraordinary’. Weak theories have little prior evidence. Their prior probabilities are low. To support such a theory would require new evidence that is, again, ‘extraordinary’.

Types of Science

Different lines of inquiry have been categorized as “hard”, “soft” and “historical” sciences. It is generally accepted that such divisions are made with respect to degrees of certainty. Philosopher and scientist Massimo Pigliucci outlined these divisions quite well in his book Nonsense on Stilts. Physics and Chemistry are considered ‘hard’ as they produce results with high degrees of certainty as experiments can be controlled with high precision. Sociology is considered ‘soft’ because it produces results that typically have lower degrees of certainty, as experiments depend on observations of people’s behavior. However, so-called soft sciences historically produce results with similar reproducibility as the so-called hard sciences.

Medical science is a mixed bag of hard, soft and everything in-between. It encompasses the high precision of physics in fields such as Radiation Oncology, and of chemistry fields such as pharmacology. Medical science includes the less “hard” fields of infectious disease and endocrinology. Medical science includes the relatively soft fields of psychology and psychiatry. All of these are science. It is all based on knowledge produced with proper applications of the scientific method.

Consensus and Paradigm Shifts

Science is a systematic method for building predictive models for the natural world. Science is also a social and human activity. Scientists are people and subject to the biases of people. Scientific knowledge should ideally be totally free of bias. As this is likely impossible, our best scientific notions are those that have been properly tested to reduce bias, reproduced, confirmed, and survived attempts at falsification by peers. Such theories and laws become very strong. Eventually, they become accepted by the community of scientists with relevant expertise. Thus, a consensus is formed.

Consensus, itself, does not necessarily make a theory valid. However, strong theories have better explanatory power than competing theories. It stands to reason that a consensus would form among scientists in support of strong theories. Such theories may achieve such wide acceptance that they become foundational. In Bayesian terms, the community would consider its prior probability to be very high. Supporters of new, competing theories have a very large burden of proof to bear.

A popular social media commentator explained scientific consensus this way:

Science is not a democracy. A consensus in science does not mean that one side has more votes than the other. A consensus indicates simply that scientists have stopped arguing among themselves. It means ideas have been tested and retested, points have been raised and refuted, and faulty hypotheses have been abandoned. It means research has narrowed to avenues which continue to make sense. It’s not a matter of lobbying the loudest for your opinion, it’s a matter of breaking under the sheer weight of compelling evidence. A scientific consensus is not an agreement that an idea is right, it’s an agreement that an idea has survived the process of science.

Twentieth century philosopher and historian, Thomas Kuhn, referred to such a consensus as a “paradigm”. Since a scientific theory is really our best available explanatory model for a phenomenon, small anomalies in the theory’s predictive power may emerge. Often, this simply results in a modification of the theory. However, if an anomaly persists in the theory’s predictive abilities, then it becomes apparent that a better theory may be needed. However, a competing theory must have all of the predictive power of the favored theory plus account for the observed anomalies. If such a theory emerges, its value may become recognized by experts. Eventually, the paradigm may give way in favor of the new theory.

Kuhn referred to this as a “paradigm shift”. He pointed out that paradigm shifts may occur later than one would expect. The scientific community is made up of people. People, even scientists, are biased toward their previously held beliefs. We do not let go of them easily or gladly (see Cognitive Dissonance). We tend to regard competing ideas/ beliefs as ‘extraordinary’. It is not until evidence for the competing theory becomes too undeniable, too extraordinary, that reasonable people finally abandon the previous model in favor of the new one.

In this way, science makes progress. Its progress is messy and non-linear. Germ theory replaced the idea of spontaneous generation. Practices such as lobotomy and bloodletting were abandoned in light of contradictory evidence and modern medical science. Science is not perfect, but eventually it is self-corrective.

Some ideas have supporters that continuously uphold them despite undeniable anomalies and despite the existence of superior theories. Such ideas are referred to as pseudosciences. They do not progress. They resist self-correction. They overly rely on confirmation and ignore falsification. Pseudoscience is the topic of the next section.


Science is a human endeavor. It is a group activity. It is competitive and moves in fits and starts. Over time, it is self-correcting. Science systematically evaluates ideas, keeps and modifies the good ones, and eliminates the bad ones. From this, we accumulate knowledge that we may use to inform our policies and decisions. Scientific knowledge is superior to other ‘ways of knowing’ because it is testable. It comes with evidence.

In science, to know something means to accept an idea in the form of a “fact”, “law” or “theory”. These 3 types of scientific knowledge inherently have degrees of uncertainty. Science’s built-in uncertainty is not a weakness. The uncertainty in science is quantifiable, and therefore allows us to understand and use the uncertainty. Science denialists may try to exploit science’s uncertainty with naive phrases like, “It’s only a theory”. In doing so, denialists inadvertently tip their hand to reveal a deep and fundamental misunderstanding of science.

** Science, by itself, is morally neutral. It produces knowledge that is reliable. It allows us to construct working models of reality that make accurate and reproducible predictions. It give us usable knowledge. The knowledge can be used for good or for bad. That is up to the user. The decision to use science to inform our actions and ethics is a philosophical one. Science gives us an idea about how the world ‘is’ to help us decide what we ‘ought’ to do.

Through this give-and-take between science and philosophy, we make progress in our understanding of the world and how we should behave.

John Byrne, M.D.

References and Links

“Robert Hazen on science “I admit, the quantum… – Lapidarium.” 2011.


“Quote by Richard P. Feynman: Scientific knowledge is a body of …” 2012.


“Talk:Bertrand Russell – Wikiquote.” 2004. <


“Not even wrong – Wikipedia, the free encyclopedia.” 2006.


“National Academy of Sciences.” 2002.


“Baconian method – Wikipedia, the free encyclopedia.”


“Hypothetico-deductive model – Wikipedia, the free encyclopedia.” 2006.


“Karl Popper – Wikipedia, the free encyclopedia.” 2004.


“Falsifiability – Wikipedia, the free encyclopedia.” 2003.


“Protoscience – Wikipedia, the free encyclopedia.” 2003.


“Ignaz Semmelweis – Wikipedia, the free encyclopedia.” 2004.


“Null hypothesis – Wikipedia, the free encyclopedia.” 2004.


“Science-Based Medicine » Prior Probability: The Dirty Little Secret of …” 2011.


“Nonsense on Stilts: How to Tell Science from Bunk, Pigliucci.” 2011.


“The Structure of Scientific Revolutions: 50th Anniversary Edition by …” 2012.