Science, Philosophy, & Reality

A Response to Sam Harris – Part 1 of 2

The Philosopher's Blueprint.Sam Harris has published his response to my winning essay in the Moral Landscape Challenge. I’d like to thank Sam for issuing the challenge and replying to my essay. I’d also like to thank Russell Blackford for all his hard work as contest judge.

In my essay, I say that Sam has “not brought questions of ethics into science’s domain.” I claim that ethics remains a “distinctly philosophical, not scientific” endeavor. My claim implies a distinction between science and philosophy, which in turn implies a difference in what we mean by “science” and “philosophy.”

In Sam’s response, he opens with the meaning of “science.” Here I reply to his remarks. In a future post, I’ll say more about the other two parts of Sam’s response, in which he talks about ethical theories and the meanings of “should” and “ought.”

CONTENTS


An empirical/conceptual distinction between science and philosophy

Sam argues that the “science” in his book’s subtitle is not science “bureaucratically construed.” Rather, it is a “domain of justified truth-claims” apprehended by “processes of thought and observation that govern all our efforts to stay in touch with reality.”

I do not assume a bureaucratic idea of science. In my essay, the science/philosophy distinction I invoke is epistemological. Science conducts  empirical investigations and makes descriptive discoveries. Philosophy pursues conceptual questions and can offer normative (which includes prescriptive) answers.

Seen as marking off areas of emphasis or specialization, the empirical/conceptual boundary can be crossed. (Bridging the descriptive-normative divide is something I’ll cover in a future post.) Experimental philosophers, like Joshua Knobe and Eddy Nahmias, cross that border when gathering data on people’s intuitions about the self, free will, or other philosophical issues. Moral psychologists, like Paul Bloom and Jonathan Haidt, cross it when designing experiments to study the development and use of moral concepts.

However, arguments and the truth-claims they aim to establish can be paradigmatically scientific or paradigmatically philosophical. For instance, Sam’s “worst possible misery for everyone” argument best fits the mold of moral philosophy: it analyzes a fundamental evaluative concept (moral “badness”) and stakes out a metaethical position (moral realism). In contrast, his argument that belief processing involves the prefrontal cortex best fits the mold of neuroscience: it defends a hypothesis that links particular mental processes to particular brain states through systematic empirical testing (even if his findings prompt, as he claims, conceptual reflection on the fact-value distinction).

As quoted above, Sam’s broad definition of science implies a difference between “thought” (conceptual) and “observation” (empirical). Elsewhere in his response, he again nods to a difference between “conceptually valid” and “empirically valid” methods for determining truth. Nevertheless, he omits any differences in epistemological emphasis or specialization that, in my view, distinguish science (empirical) from philosophy (conceptual).

Sam Harris’s three defenses of his broad definition of “science”

Sam makes at least three attempts to obviate criticisms of his capacious conception of science. Below I summarize and respond to each attempt.

Science is a synonym for the unity of knowledge

Sam claims that science, broadly construed, refers to “the unity of knowledge.” Describing this unity, he writes:

[T]he boundaries between [academic] disciplines are mere conventions…[W]e inhabit a single epistemic sphere in which to form true beliefs about the world.

To help illustrate this idea, Sam offers an analogy. He compares ‘truth boundaries’ between different fields of study to political boundaries on maps. There’s no real boundary between some neighboring states—just shared terrain, such as a desert or a forest. Sam says there’s likewise no real boundary between neighboring disciplines like “physics and…biology.” When Sam speaks of “non-contiguous disciplines,” such as “physics and sociology,” the analogy changes. He doesn’t compare their terrain (natural vs. social) to deserts or forests. Instead, he compares it to “different colors on the map” for different states. He says he’s convinced the significance of these colors and, by analogy, the difference in facts of distant disciplines is an “illusion.”

But it remains unclear why we should be convinced. I’m pointing to the line between the sciences (natural or social) and the humanities, particularly philosophy. The territories I identify are empirical and conceptual. If you’re deep in the former, you’re doing science. If you’re deep in the latter, you’re doing philosophy. And if you travel between the two, you cross a border that isn’t merely conventional.

Even if there is a unity of knowledge across academic disciplines, it remains unclear why we should call that unity “science.” After all, it would be a unity of knowledge from the sciences plus philosophy and the other humanities.

There’s “no telling” what questions science will address

Sam claims we cannot predict where and how seemingly non-scientific projects will become “entangled” with the sciences. As Sam suggests, sometimes questions pursued by, say, historians (who asked whether the Shroud of Turin is from the first century) will come to be pursued by, say, chemists (who answered using radiocarbon dating).

Even if some questions first raised outside of particular scientific disciplines can unforeseeably fall within their domain, it doesn’t follow that all questions about reality should be (preemptively?) placed within science’s purview. Sam nonetheless endorses the conclusion that science  is the sole arbiter of reality.

It’s an irrelevant semantic choice

Despite giving what look like earnest arguments for his broad definition of “science,” Sam seems to dismiss this dispute as semantic quibbling. He writes:

The whole point of The Moral Landscape was to argue for the existence of moral truths—and to insist that they are every bit as real as the truths of physics. If readers want to concede that point without calling the acquisition of such truths a “science,” that’s a semantic choice that has no bearing on my argument.

The first sentence reveals that Sam’s book has a philosophical objective. He intends to defend moral naturalism. In Sam’s words (taken from his “central argument”), moral naturalism makes two claims:

  1. “questions of morality and values must have right and wrong answers”
  2. the answers in (i) “depend on… natural phenomena, fully constrained by the laws of the universe”

However, the “depend on” in (ii) doesn’t quite capture the connection Sam sees between an objective morality and the natural world. Sam argues for analytic moral naturalism: statements of moral fact conceptually reduce to statements of natural fact. Put another way, the definitions of moral terms, such as “good” and “bad,” are given by descriptions of the natural world. For Sam, the relevant descriptions refer to peaks and valleys of conscious experience on the “moral landscape.”

If analytic moral naturalism is correct, then calling the “acquisition of [moral] truths a ‘science’” is not a mere semantic choice that’s irrelevant to Sam’s argument. Rather, it’s the conclusion. Like uncovering the facts of physics, finding the truths of morality will “fall within the purview of science (in principle, if not in practice).” Sam uses those words to complete the quote in (i). And that quote is, in fact, the conclusion of Sam’s argument for the existence of moral truths.

I don’t object to calling normative ethics “science” if analytic moral naturalism is correct. But I wouldn’t call arguing for any of the following claims “science.”

  • Analytic moral naturalism is correct.
  • Well-being is the only thing of intrinsic moral value; it encompasses all other moral values.
  • Morally right actions maximize collective welfare.

Sam argues for all of the above. In fact, these are the most crucial claims he makes in The Moral Landscape. And they are all philosophical claims that elude scientific (read: empirical) demonstration. Science cannot “determine human values” as Sam envisions unless philosophy establishes his core claims—especially the second claim that all human values boil down to well-being.

Contrary to what Sam seems to suggest, I’m not thinking about the architecture of universities or any other aspect of science “bureaucratically construed.” I’m thinking about the architecture of reality, which I consider integral to “the nature of reality,” Sam’s stated interest. If the work of reconstructing reality is ever complete, the edifice that emerges may well stand, foundation to spire, within the natural world—science’s domain, ontologically speaking. Epistemologically speaking, the blueprint will trace a complex composite of concepts spanning various levels of analysis (e.g., moral, social, psychological, biological, physical, natural). The coherence of those concepts and, where possible, the bridging of those levels of analysis will depend crucially on work best called philosophy, no matter where on campus it is conducted.

The foundation of science

In answer to my renewal of the Value Problem objection, Sam renews his response to critics that science cannot derive any judgment, moral or otherwise, “solely from a scientific description of the world.” His emphasis, however, has shifted from the necessity of foundational values to that of foundational intuitions, with our basic notions about causality now in the mix. I don’t protest this shift. Values and intuitions may overlap, and Sam’s appeal to science’s dependence on the latter is found in his book. In my essay, I try to show that (quoting Sam) “other branches of science are self-justifying in a way that a science of morality could never be.” I may have given readers the impression that Sam thinks this: like other sciences, a science of morality will be foundationally self-justifying. Sam makes clear in his response that he in fact thinks this: no science is foundationally self-justifying and thus no science of morality will be.

I’m uncommitted regarding whether “the epistemic values of science are ‘self-justifying’.” In my essay, I simply try to paint a coherent picture of what it would mean to convince Sam (or others) of that claim. When I write that “science [can’t] appeal to health to defend health’s value, as it would appeal to logic to defend logic’s value,” I don’t say the latter appeal will be successful. Perhaps it’s viciously circular. But logic is both a concept and a tool for justifying concepts, so it can be coherently, even if unsuccessfully, applied to its own defense. The same can’t be said for health. In any case, I don’t intend to hold Sam’s science of morality to “a standard of self-justification that no branch of science can meet.”

I also don’t deny that epistemic axioms “do far more than motivate scientific inquiry.” I say those axioms “merely motivate scientific inquiry and frame its development.” That is, they perform merely two tasks. To frame science’s development is to, as I say, “direct science to favor theories that are logically consistent, empirically supported, and so on.” To me, that idea sounds like Sam’s claim that epistemic axioms “determine what we find reasonable—or even intelligible—at every stage of [scientific] inquiry.”

In my essay, I criticize a comparison Sam makes in his book’s afterword. Below I’ve updated that comparison to include intuitions. I’ve also restated my criticism.

Sam’s science of morality assumes parity between

  1. valuing truth, logic, evidence, and health; intuiting the contours of rationality and causality
  2. valuing well-being—such that all moral values reduce to well-being

I challenge that assumed parity. The analogy fails. Well-being’s value, as conceived in (b), doesn’t have the axiomatic status that Sam ascribes to the values and intuitions in (a). What’s more, a science of morality, I say, requires parity between

  1. scientific theories—such as physicists’ competing ‘theories of everything physical’
  2. ethical theories—which I’ll call philosophers’ competing ‘theories of everything moral’

If physics were to presuppose that string theory is correct, it would commit the deeply unscientific sin of asserting a fundamental, revolutionary finding that it hadn’t actually made. If Sam assumes that science will elucidate moral reality by identifying the correct ethical theory, his science of morality commits that same sin when it presupposes that welfare maximizing consequentialism is correct.

Sam need not accept the parity I propose between (c) and (d). He could reply that moral philosophy, not science, identifies the correct ethical theory. It would just be the case that, in his view, moral philosophy—as guided by intuitions he finds “non-negotiable”—has demonstrated that welfare maximizing consequentialism is correct.

Indeed, perhaps The Moral Landscape is best interpreted as follows. For Sam, the philosophy is done and the science (especially the neuroscience) has already begun. Debate in metaethics can be closed: analytic moral naturalism has prevailed. Debate in normative ethics can likewise cease: welfare maximizing consequentialism has bested its deontological, aretaic (i.e., virtue ethics), and fellow consequentialist competitors. Just as natural philosophy long ago became the natural sciences, moral philosophy now becomes the moral sciences. More precisely, it becomes the science of well-being.

In my next post, I’ll see what I can do to keep moral philosophy in your local university’s budget.

Comments 17

  • At last, someone worthy to redeem the good name of philosophy! That dark cloud that has hung over it for so long now gives way to a bright ray of sunshine. The lines of demarcation are drawn, or I should say re-established with an eloquent voice of clarity. I know all this sounds a bit over the top but I can’t emphasize enough the sheer joy, dare I say “religious experience”, after spending so long in the quagmire that philosophy had become, to be able to breath clean air once more.

    • Thanks for the high praise, Keith. I don’t think the clouds have cleared just yet, though. But I’ll do what I can to help illuminate the work philosophy can do.

  • The issue of self-justification seems conceptually confused to, so I’m afraid your approach may have given Harris too much wiggle room. I think the easiest way to reveal the problem with Harris’ approach to morality is to directly attack his interpretation of “ought” and “should,” so I look forward to your next post, which you say will address that issue. As a possible resource, check out my response to Harris’ Moral Landscape Challenge. It was not entered into the competition, so I didn’t actually lose. 🙂 In any case, it’s short and I think you might find it interesting and useful.

    • Whether or not it’s conceptually confused (and it’s at least unwieldy), the issue of self-justification can be safely set aside. That is, we can argue against the axiomatic status Sam gives to valuing well-being without arguing for the self-justification of the values and intuitions to which Sam compares well-being’s value.

      By the way, thanks for the link to your piece. I’ll check it out.

  • “Science cannot “determine human values” as Sam envisions unless philosophy establishes his core claims.”

    Yes! My favourite line from this piece. I’m also glad to see you leave room for moral naturalism as that’s the tack I took in my Sam Harris response—providing the objective natural basis for morality (http://is.gd/IzLP3e). Sam just took morality (acting well), and defined it as acting for well-being. Acting well by acting for well-being without defining what “well” is? Well, that’s pretty darn circular I’d say.

    Looking forward to your next piece.

    • Thanks, Ed.

      I’m open to moral naturalism, as I lean toward both moral realism and metaphysical naturalism. However, I’m still not settled about the reduction from moral truths to natural facts.

      Thanks for the link to your piece. Part 2 of my response to Sam will be up soon.

  • I think there’s an interesting point to be made here:

    “Science conducts empirical investigations and makes descriptive discoveries. Philosophy pursues conceptual questions and can offer normative (which includes prescriptive) answers.”

    Are you saying that science does not pursue conceptual questions, and does not provide normative answers to questions? Or are you saying that, when scientists do that kind of work, they are doing philosophy, not science?

    • I want to say that professional scientists (e.g., physicists, psychologists) venture into philosophical territory when they’re developing ideas that inform their empirical-descriptive work but resist verification by it.

      For instance, the psychologist B.F. Skinner argued for the doctrine of behaviorism, which states that psychologists study observable behavior, not unobservable mental events. In fact, Skinner maintained that mentalistic concepts (e.g., belief, desire) can in principle be purged from psychological theorizing, leaving just behavioral concepts. Skinner looks to be doing some philosophy of mind here. Specifically, he is proposing the elimination of one set of concepts about the mind in favor of another.

      • That seems quite consistent, thank you.

        I’ll lay my cards on the table from the start, my intuition is that your distinction between conceptual work and empirical work is valid but I disagree about how to draw the line between science and philosophy.

        To me (as someone who is not a philosopher but has an interest), it seems that the way we use the word philosophy is to describe the activity of building theories of reality, and we use the word science to mean the collection of data about reality.

        For example, in the 70’s the cognitive psychologists were building different theories of memory processing, and two of the main theories were the levels of processing theory (the more you think about things, the stronger the memory trace that is left) and the multi-store models of memory (there are different types of memory; short-term, long-term, etc). They took the data from scientific experimentation with regard to memory, and used empirical experiments the same way that (I think) philosophers use thought experiments, that is, to assist people’s intuitions in understanding the theory. Using these models, they were able to make predictions about what they should expect to be the case if they were correct or incorrect. My view of how philosophers use thought experiments is largely based on Dennett’s book ‘Consciousness Explained’, so I realise that my view is not particularly sophisticated, I’m interested in being corrected.

        Now, you said that philosophy involves building models that resist empirical verification, but I don’t think the example of Skinner’s behaviourism is a good example. As it turns out, the theory was largely falsified through experimentation, for example in Chomsky’s examples of language being partially innate (as opposed to completely conditioned as behaviourism would have then predicted) and also in the Bobo doll experimentshowing that we should assume that cognitive processes were occurring unrelated to conditioning. I would agree that the building of the behaviourist theoretical framework was a philosophical action, but I disagree that philosophy is by definition resistant to empirical verification.

        For this reason I agree with Harris on not making a distinction between philosophy and science in morality, though I think there is a valid distinction to be made between theorizing and data collection that the words could be used to describe. In an ideal world, of course…

        • I agree that philosophers do some theory building, as you say. I think scientists also do theory building, as you illustrate with cognitive psychologists’ theories of memory. But if there’s a difference between the theories they build, it will have to be, in my view, that philosophers don’t build empirical theories—that is, ones that are subject to (dis)confirmation through the collection of empirical data.

          You have a point that behaviorism is a problematic example of a non-empirical theory. Skinner was committed to the claim that the sources of behavior are external, not internal—that is, not mental but environmental. This claim entails that learning the grammar of one’s native language will be explicable in terms of environmental input. But, arguably, to a substantial extent it’s not, and empirical evidence can help make the case. (However, I wonder: could a “poverty of the stimulus” argument succeed on just a priori grounds?)

          So perhaps the easiest example will be theories about the nature of science and philosophy! For instance, the doctrine of logical positivism (aka, logical empiricism) proposed the verifiability principle: a meaningful statement must be either empirically verifiable or analytically true. This principle was believed to render much of traditional philosophy—including metaphysics and ethics—meaningless. Meanwhile, science gets a big boost, given that empirical verification is its thing. However, the verifiability principle appears to render itself meaningless, as it looks to be neither verifiable nor analytic. If so, then the doctrine of logical positivism (which aimed to subsume philosophy within science) stands or falls on conceptual (hence philosophical) grounds.

          • Hmm, see I’d be interested in seeing a philosophical theory that is both useful and not subject to empirical verification. I suspect that truly non-empirical theories would be akin to epiphenomenalism.

            I don’t think an a priori argument would work for the poverty of stimulus argument, it relies on empirical information like much information a child receives about a language and how they are typically taught language, and the types of grammatical errors they make when growing up. It is entirely plausible that some kinds of brains could be blank slates that just learn through conditioning, the brains of humans just don’t seem to work like that (nor do I know of other types of brain that are like this).

            With regard to the verifiability problem of logical positivism, would you consider the principle of non-contradiction to be a philosophical thinking tool? Is logic within the domain of philosophy, not science?

          • I’d be interested in seeing a philosophical theory that is both useful and not subject to empirical verification. I suspect that truly non-empirical theories would be akin to epiphenomenalism.

            Even if the principles that compose a theory aren’t empirically verifiable, they can still have empirical application—that is, have some use or effect in the “real world,” as one might say. For instance, the rationalist accounts of ethics that Harris rejects (like Kant’s deontology) rely on a priori principles that resist empirical verification but are nonetheless action guiding. We’d use them in combination with empirically informed judgments about the circumstances in which we act. But, admittedly, this may not the best example if you’re inclined toward Harris’s ethical theory.

            About the pos argument, I suppose I’m thinking that there’s an element of mathematical necessity to its conclusion. In any case, yes, absent observation about the language learning environment, the pos argument appears unable to establish nativism about language.

            To answer your final questions, laws of logic must be used by both scientists and philosophers. In fact, I wouldn’t call the use of logic an exclusively philosophical enterprise by any means (in case that’s a worry). However, systematizing and formalizing logic has tended to be a philosophical endeavor. For instance, categorical and propositional syllogisms were first developed (or, in a sense, discovered) by Ancient Greek philosophers. But then mathematicians really got in on things in the 19th and early 20th century.

  • Brilliant keep talking guys, this is good stuff, my FAITH is renewed in philosophy!

  • Congrats Ryan on winning the TML Challenge. A good essay, that I think hits on the essence of the problem—even if SH doesn’t agree. I was disappointed that there wasn’t an extended conversation, though I’m not surprised that things degenerated into “hairsplitting and backtracking.” My only concern with your essay was that the technical style, though philosophically reasonable and something Blackford obviously favors, allows Harris too much opportunity to muddy the waters by piling on the epicycles of philosophical complexity to his own theory. I think Blackford should have recognized that that if technical was what SH needed to hear, then he should have been able to convince SH in his own exchange with him. I’m just making an observation and am not trying to take anything away from your win – enjoy the $2000!

    • Thanks, JJ. Making the exchange accessible to a non-philosophical audience was a concern. To that end, I did try to avoid terminology not found in Harris’s book (though you’ll have to dig in the footnotes to find mention of “epistemic values”).

  • I’m not convinced by your science vs philosophy distinctions. “worst possible misery for everyone” can be as scientific as anything else. Within the bounds of ordinary cognitive science you could specify misery as a set of brain event types with a scale of severity. You can also specify sets of brains containing misery events and the social relations between the brains. There’s no need to ghettoize such phenomena as “evaluative”, “metaethical”, etc. Seems you’re just fighting another turf war against an emerging science. We know how those always end: dusty bookshelves.

    I don’t see any reason, and you don’t give any, to believe your claim that bridging the various levels of analysis used in science would fall outside of science. Science bridges different parts of science. You don’t need to invoke your “philosophy” bogeyman for that. The relationship between chemistry and biology is found within those fields. You don’t ask a philosopher. Morality is just another system that emerges from sets of biological organisms interacting. Understanding that system requires some fancy computational/mathematical models that we haven’t devised yet, but when we do, it will be science as normal, not philosophy.

  • […] Harris has written a response to Born’s essay.  Born has, likewise, responded to Harris (Part 1 and Part […]