Sam Harris and the Demarcation Problem

6a00e554e8195d883301a73d8ae293970dby Paul So

Sam Harris is known for many things, from being one of the leading figures of the New Atheist movement to a controversial critic of Islam. he is also known for arguing that science can provide answers to questions regarding morality [1]. For him, morality is within the domain of science.

How is this possible, exactly? After all, science deals with facts, not values. Harris proposes that the term science is far more inclusive than we normally understand. There is no fundamental distinction, for instance, between a scientist working in a laboratory and a plumber identifying problems in a plumbing system. The distinction between them is merely conventional, because what really counts is that doing science means using reason and observation. As long as a given domain can be the subject of reasoned inquiry and observation, it belongs to the broader domain of science.

Here is how Harris himself puts it in his essay responding to Ryan Born’s critique [2,3]:

“For practical reasons, it is often necessary to draw boundaries between academic disciplines, but physicists, chemists, biologists, and psychologists rely on the same processes of thought and observation that govern all our efforts to stay in touch with reality.”


Many people think about science primarily in terms of academic titles, budgets, and architecture, and not in terms of the logical and empirical intuitions that allow us to form justified beliefs about the world.” 

As well as:

“I am, in essence, defending the unity of knowledge — the idea that the boundaries between disciplines are mere conventions and that we inhabit a single epistemic sphere in which to form true beliefs about the world.”

It seems from the above that Harris thinks science is just the application of reason and observation in order to arrive at justified true beliefs about the world. To be more precise, what Harris would likely say is that science is conceived as consisting in applying reason and observation with the intention to acquire justified true beliefs (otherwise, every time a scientific notion turns out to be false we would have to conclude that it wasn’t science to begin with).

As long as we use reason and observation with proper epistemic intentions, we are doing science regardless of whether or not we really do acquire true beliefs. Even if we found out that some of our beliefs are false, we can always try to replace them with better justified ones. While this definition too is not without problems, I will assume this is what Harris has in mind. I’ll argue, however, that if science is conceived as just applying reason and observation with an intention to acquire justified true beliefs, this leads to one of the best known problems in the philosophy of science: the demarcation problem.

The demarcation problem is the problem of how we differentiate science from pseudoscience in principle. In other words, the demarcation problem consists in finding some principles, criteria, reasons, or conditions to place something like astronomy under “science” and place its counterpart astrology as a pseudoscience. However, for every proposed claim about what distinguishes science from pseudoscience, there’s a counter-example. For example, Karl Popper proposed that falsification is the criterion for distinguishing science from pseudoscience. If any set of claims or theory is falsifiable, it belongs to the domain of science. But if any set of claims or theory is unfalsifiable, it belongs to the domain of pseudoscience. However, this criterion is too strict because some untestable scientific claims ranging from string theory to the many-worlds interpretation are not considered pseudoscience.

How is the problem of demarcation relevant to Harris’ definition of science? If science is any activity that relies on reason and observation (or empirical and logical intuitions) with the intention to produce justified true belief, then the concept includes many things that are considered to be pseudoscience or non-science. Consider three examples.

First, phrenology. Phrenology is now relegated to pseudoscience, but during the 19th century many people took it seriously. Phrenology claims that personalities, emotions, talents, and such are caused by the activity of very specific regions of the brain. The theory of Phrenology was developed by Franz Joseph Gall, on the basis of his observations of the size of many skulls. Harris’ definition of science seems to force him to accept that phrenology is in fact a science.

Second, Intelligent Design (ID). Many people like to ridicule proponents of ID as mindless buffoons, but in fact the public figures of ID like Stephen C. Meyers and Michael Behe are well-educated and thoughtful people. This doesn’t mean that their claims are true. After all, it’s possible to be well-educated, thoughtful, and yet fundamentally misguided. But ID proponents like Stephen C. Meyers do in fact use reason and observation to support their claim. They give arguments and provide what they think of as empirical evidence for their conclusion that there must be an intelligent designer. In effect, they are doing science according to Harris’ definition.

As a side note, someone could object that ID proponents are using reason and observation too poorly for what they do to be considered science. Moreover, what they are doing goes against the established body of knowledge. Yet, Harris’ definition of science does not really include any qualification concerning the quality of using reason and observation. Harris could propose to amend his definition to say that reason and observation need to be used well. I shall address this later. As to the second point, going against the body of established knowledge may seem irrational, but we want to be careful because many scientists who initiated a breakthrough were going against established the then accepted body of knowledge. Albert Einstein’s General Relativity went against Newtonian Mechanics, which was an established body of theoretical knowledge. However, we certainly don’t want to say that Einstein was being irrational.

Third, consider Natural Theology. Regardless of what one may think of Natural Theology, we can all agree that it is not a science. However, natural theologians use observation and reason to support their claim that God exists. One notable example is the fine tuning argument. Natural theologians observe that the values of the cosmological constants are conducive to the existence of life, and they make an inference to the best explanation (at least in their view) that God is responsible for so structuring the universe. Whether or not this is a convincing argument, natural theologians are indeed using both observation and reason to support their claim. According to Harris’ definition of science, Natural Theology is therefore a science.

What I’m trying to argue by way of these counterexamples is that Harris’ conception of science is far too broad. It readily includes a number of notions that most of us wouldn’t consider to be within the domain of science, and reasonably so. In fact, it seems to include things that are considered to be downright pseudoscience, or theology. It should therefore be apparent that Harris’ definition of science is not very helpful, as it exacerbates the demarcation problem.

Harris could reply by arguing that he wants to make a distinction between a rigorous and reliable use of reason and observation and a loose and unreliable use of reason and observation. Anything that counts as science involves a rigorous and reliable use of reason and observation, whereas pseudoscience involves a loose and unreliable use of reason and observation, although both have the intention to produce a justified true belief. With this new distinction, Harris could exclude ID, Phrenology, and Natural Theology from the domain of science because they involve a very sloppy and unreliable use of reason and observation.

However, even Harris’ improved definition wouldn’t work. Even if it succeeds in excluding ID, Phrenology, and Natural Theology, it ends up excluding a lot of science. There are many scientific works that are not very rigorous and reliable. For example, more often than most people realize, peer review journals tend to publish scientific works afflicted by serious methodological problems.

One also has to consider the kind of scientific work at the frontier of knowledge. A lot of this works will turn out to be mistaken, because it is dealing with something that is barely within the grasp of science. For example, consider some cutting edge work on neuroscience. Despite the fact that neuroscience has made enormous progress of late, there is of course still a lot that neuroscientists do not know. Application of the scientific method in this domain began rather poorly (not rigorously, and somewhat unreliably), but eventually improved, and continues to improve. Still, even according to Harris’ augmented definition, neuroscience done poorly is not within the domain of science.

Harris may, of course, continue to modify and improve his definition, but he also wants to maintain it as broad as possible, in order to include ethics within the domain of science. This is a very difficult, if not impossible, challenge. What we have seen so far is that he has to narrow down his definition in order to avoid embarrassing counter examples. But if he is forced to continue on this path, there’s a potential problem that his eventual definition will end up either being beyond recognition and familiarity, or fail in its stated purpose to include branches of philosophy such as ethics.

Some readers may at this point conclude that this is merely a semantic issue. In an important sense,  they are correct. After all, Harris provides a definition of science, and I am disputing it. This is a discussion about the meaning of words, that is, about semantics. But contra popular understanding, semantic issue aren’t pointless. On the contrary, they are quite instructive. Mine is a cautionary tale on what happens when one broadens the meaning of an important word too much, leading straight into clearly unintended and perhaps even embarrassing consequences. One simply has to be careful with how one uses words, especially when one’s entire argument depends on it. Despite being a good writer, Harris, it turns out, is not careful with his words.


Paul So is a graduate student studying for the Master’s program in the Philosophy Department at Texas Tech University. His main focus is Philosophy of Mind.

[1] The Moral Landscape: How Science Can Determine Human Values, by S. Harris, Free Press, 2010.

[2] Clarifying the Moral Landscape, by S. Harris.

[3] The Moral Landscape Challenge: The Winning Essay, by R. Born.


Categories: essay

Tags: , , ,

103 replies

  1. Hi Coel
    I realize you are out of comments, so I suppose I hope you will perhaps consider these. Also if anyone shares Coel’s views (or wants to play devil’s advocate) I would be interested on this or the issues I raised in my last comment regarding beliefs formed based on testimony. I happen to think allot of our beliefs are formed based on testimony.

    As far as what Coel said, I agree it’s true that science can inform our opinion about what happened in the past even for a one off event. The case you bring up about the extinction is a good one. Another would be humans evolving from other species. We don’t need to repeat the exact thing in science to have science inform us about what happened.

    But I think it’s still possible to say that the actual science does involve repeated testing – either in the lab or in the field – were variables are isolated and predictions come true or not. It’s true we don’t have another mass extinction or see another human evolve, but we do have repeated experiments and findings that show how that could have happened.

    But my last post got to another way that our knowledge of history depends on things that are not properly science. It often relies on people simply telling us what happened.

    The cases you mention are not really history they are pre-history. Science is all we have to try to figure out what may have happened. The line between history and prehistory is usually drawn by whether or not we have written records of what happened. In other words, can we take anyone’s word for what happened? Historians analyze these records to determine what likely happened.

    One historical question might be whether Martin Luther said “here I stand” at the Diet or Worms. Science might tell us it was possible that he said it. But science doesn’t really help us with whether he, in fact, said it. Science cannot really help us determine if he, in fact, said anything at all. If we want to believe Martin Luther said anything at all at the Diet of Worms, we need to address people’s claims that he was there and he spoke.

    I am not sure whether he said “here I stand” at the Diet of Worms, but I do believe he said *something* at the diet of worms. And my belief is due to people telling me he spoke there. Yes they are telling me in writing but it is still based on their “say so” as are so many things in history. Is my forming a belief that Martin Luther spoke at the Diet of Worms, based on this “say so” properly considered science? If you agree it is not, then it seems you either have to admit we know things outside of science, or admit we can know *very* little about what happened in the past.


  2. Coel said:
    Objective morals (moral realism) need to be entirely independent of anyone’s feelings. My sole point here is that such a notion is nonsensical, because morals *are* feelings that evolution has programmed into us.

    I read those words with disbelief. The sociologist, Christian Smith, reported that young adults in transition from childhood describe such elementary moral judgements. A typical immature young adult would say ““I would do what I thought made me happy or how I felt. I have no other way of knowing what to do but how I internally feel.”” (

    Here are some feelings, do they describe you? Hate, vengefulness, lust, avarice, envy, pride, greed, wrath, covetousness, sloth, etc. I am sure you will deny this and instead you will claim nobler feelings. But on what grounds will you do that? Why choose some feelings over other feelings? We all do that and the moment we choose one set of feelings over another we have started to practise moral thinking.

    Only the most immature let their feelings dictate their behaviour. As we grow into maturity we learn three things:
    1. Self control. We restrain or delay acting on our feelings in order to attain a larger goal. The two marshmallow test is a lovely childhood example of that.
    2. Morality. We choose to deny certain feelings/needs and choose instead to act on other feelings, according to an ethical framework.
    3. Altruism. We sacrifice our feelings/needs for a greater goal.
    We learn restraint, morality and sacrifice.

    All three have this in common:
    1. a reflective awareness and internal examination of the feeling.
    2. the feeling is subjected to an evaluative process.
    3. there is a disciplined readiness to control the way we act on the feeling.
    We practise awareness, evaluation and control.

    It is this reflective awareness coupled with an evaluative process, directed at other than needs centred goals that makes up moral thinking. You define morals as feelings. But feelings cannot be morals. Morals are the evaluative process that we apply to the combination of 1) the facts, 2) our feelings, 3) our needs and 4) our goals.

    it is when we engage in conscious evaluation of the quadrant of facts, feelings, needs and goals that we start to practise moral thinking. But it is more than that. I do all four things in the planning of my running programme but that is not moral thinking. It becomes moral thinking when I add an essential fifth ingredient, that of goodness and rightness.

    Society has a shared concept of rightness and goodness. We use this fifth concept to guide the evaluative process as we apply it to the facts, feelings, needs and goals. That finally is moral thinking.

    Do you begin to see how far your primitive concept of feelings is from the complete process of moral thinking? It is typical pre-teen behaviour to act according to feelings. As we grow into maturity we learn to replace feelings with a considered evaluative process that is guided by our shared concepts of goodness and rightness.


  3. Joe – re testimony, the historical sciences and history. I suppose it depends on whether you think anthropology, sociology and psychology are sciences or not, and whether we can bring insights to bear from the social sciences onto any particular historical question. Here’s a lovely scientistic bit from

    Martin “Towards a scientific history of religions” (Ch 2 of “Theorizing Religions Past: Archaeology, History, and Cognition”) via Google Books,

    For historiography to be a reliable way of knowing, then, its methods must at least approach the standards for verification that exist within science, especially the nonexperimental “historical” sciences like astronomy, geology, paleontology and evolutionary biology…historiographic interpretation must be subject to certain general standards and tests – of human behaviour, of logical antecedents and consequences, of statistical or mass trends…

    There is or was something called psychohistory, and Erik Erikson (the adolescent psychiatrist) did write Young Man Luther: A Study in Psychoanalysis and History, where he tried to use contemporaneous observations of human behaviour and his ideas about “the crisis of identity” to interpret Luther’s life and works. I’m guessing he would suggest Luther would have said something like that (or if he hadn’t, he later remembered saying it), because it would fit his theoretical model nicely.


%d bloggers like this: