Here's the quote from Shulevitz:
Faced with these dilemmas, philosophers concluded that there was only one way to tell religion from other apparent aberrations. Religion, they said, is what makes people feel religious. It puts them in touch with the Infinite or with what Freud called "the oceanic feeling." Defining religion as an experience, however, puts us back into the old Kantian world in which science goes here and religion goes there. An experiential definition of religion renders it impervious to empirical observation. You can never prove that someone feels religious, so you can never prove that something is a religion. The philosopher Søren Kierkegaard took this idea even further, saying that the nonreligious could never understand the truth of religion, because religious experience could be understood only subjectively, not objectively.This is, in point of fact, precisely what Kierkegaard did not do.
"Truth is subjective," Kierkegaard famously wrote, but by that he did not mean that truth is mere sophistry, that what is true for me may or may not be true for you. He was following, in fact, in Hume's footsteps. Which is where we have to go, and do so by way of Kant.
Kant, as Shulevitz notes (rather reductively, but that's unavoidable in this cirumstance) divided the world into pure reason and practical reason. He was responding to Hume, who seemingly left reason in such tatters it could not recover. The problem has been ever since that some think Kant put Humpty Dumpty together again, and some understand he just invented a new Humpty Dumpty, a sort of hologram Humpty Dumpty, if you will. Dennett seems to be of the former category, the attitude that prevails (to be perfectly reductive and generalizing about it) in Anglo-American philosophical circles (and the reason they are usually dismissive of existentialism and other schools of phenomenology, which is primarily associated with "Continental" philosophers, primarily, most recently, Sartre, Heidegger, and Derrida).
So the lines are (very broadly) drawn. Now back to Hume.
Hume was the last great Empiricist, part of the "trinity" of great British empiricists, the other two being Locke and Berkeley. Empiricism (and it's influence on Anglo-American philsophers cannot be overstated) argued, contra Plato, that all knowledge comes only from sensory impressions. For Locke, this meant the mind was a "tabula rasa," a blank slate, upon which our experiences, via memory and the senses, wrote our knowledge (Locke is the reason we are so fascinated with memory and personal identity today. You can draw a straight line from Locke to Philip K. Dick, and to the movies "Memento" and "Batman Begins;" a line that obviously runs through Freud, too.)
Hume took this reasoning to its logical conclusion, arguing, among other things, that since he could not observe a "self" which in turn observed the world the senses told the "self" about, he had to conclude (as Shulevitz says Dennett does) that the "self" was an illusion (the validity of that we can discuss anon). But he also undid causation, and did so by developing a method of determining, quite rigorously, what could be known.
He did so by distinguishing two types of ideas: analytic, and synthetic. In briefest summation, synthetic ideas are those known through experience (more technically, information known a posteriori, not a priori). And example is the knowledge that a given stone is heavy. Analytic ideas are known a prior (a bow to Socrates), but are purely about relations of ideas; an example is 2 + 2 = 4. Synthetic ideas are encountered in the world, and so there are consequences to them; however, they don't reveal any truth. Analytic statements are true, but their truths are irrelevant.
Remember, these are the two kinds of statements we can make: they all divide into analytic, or synthetic, according to Hume; and while the argument over the precise nature of the divide rages in modern philosophy, the importance of the division to empiricism is not really questioned.
But Hume went one step further with his emphasis on observation. He annihilated the "self" by arguing he could never observe a self observing sensory perceptions, and therefore the "self" was just an illusion, a concept arising from the activity of perception itself, which activity, Hume argued, was all that could be known. Well, from that reduction of observation, he went to causation itself. First Hume argued that causation was a synthetic judgment; we know it only from experience, and our reasoning about it is always a posteriori. That is, we determine causation only after the fact. This much is sound "scientific" reasoning. But, says Hume, just as we can never observe a "self" observing sensory impressions, we can never observe a precise connection between the cause and the effect. Based on experience (which is a posterirori), we assume there is one. But that assumption is not proof. We cannot, actually, prove causation; we can only assert it, and that assertion is not enough. Minutely as we might observe the cue ball strike the 8 ball, we cannot say before the fact what will happen before the balls collide, especially if we are seeing it for only the first time. What we can say is what we expect to happen, but that expectation undoes our observation, because the connection between one event (moving cue ball, stationary 8 ball) and the next (moving 8 ball due, says physics, to a transfer of energy) cannot be shown. It can only be assumed, based on experience.
Must the cue ball move the 8 ball? Experience says it will. But how do we know? We must run the experiment. This, of course, is where science would seem to grind to a halt. The experiment says that, under certain conditions, certain results follow. But until we run the experiment, we don't know that is true. Once we run it, will it be true again? We'll have to run the experiment. Rather like Zeno's paradox of the turtle and Achilles, our observation is reduced to a nullity. We cannot, in other words, soundly predict what will happen. We can only, at best, say what we expect to happen.
And in that state of affairs we cannot validly predicate causation.
Kant reestablishes the validity of observation (especially our observation of causation) by linking our perceptions to reality through ideas (hence the name of Kantian school of philosophy, Idealism). He resorts, essentially, to Plato, to defeat the non-Platonic empiricism. To greatly reduce his subtle and complex argument, the crux of the matter is that while we cannot know things in themselves (the Ding an Sich for Kant, and there Hume remains triumphant) we can know the ideas related to those things, and for us, those ideas are reality, so the attempt to separate them from the things in themselves (the ding an sich) is ultimately bootless, because we only talk about the ideas anyway. Might as well accept them as real. On that split arises modern philosophy, and the Enlightenment, with it's assertion that reason is superior to belief, because the products of reasoning can be empirically established. But empirically established by idealism, not by direct observation (Kant and Hume both seal us away from the things in themselves). Kant, in other words, sort of gets folded back in, even as the staunchest materialists (like Dennett) try to reach around idealism and appeal to empiricism (if they grasp Hume fully, they are left with with his conclusion, summed up among philosophy students as: "No matter; never mind." Not a position entirely useful to a materialist.) .
This is also where modern philosophy turns toward language, something that becomes an abiding interst of philosophers as different as Austin, Wittgenstein, and (again) Derrida. All Hume really leaves philosophy with, in the end, is language. We cannot really establish anything (not without being Kantian Idealists), and so we have to talk about what we are talking about. Philosophy turns, in other words, on it's own tools: language.
Now, what does Hume have to do with Kierkegaard?
Kierkegaard is both post-Hume and post-Kant, and following somewhat on Descartes, becomes more concerned with the existence of the individual than with the "big idea" of the nature of existence. Hume has already established, at one end, that this "big idea" is at least very difficult, if not impossible, to discuss in philosophical terminology (which would lead to the intense focus of some philosophers on language itself), and Kant at the other had re-established our ability to speak of such things by appealing to trascendent ideals much larger than even humanity itself. That lead philosophy to the rarified atmosphere of Hegel's Idealism, which Kierkegaard was reacting against (in no small part because there is not much room for either individuality, or the God of Abraham active in human affairs, in Hegel's philosophy of history; at least not enough to satisfy Kierkegaard). But where Hume said we cannot speak of truth because we can never identify it objectively, Kierkegaard turned the coin over and said: well, then, truth is subjective.
But then we are referring, of course, to two different kinds of truth. The problem with Hume's definition of the self, for example, is that it leaves one (in our day, if not Hume's 18th century) with the image of a TV screen receiving pictures (sense impressions) in an empty room, yet somehow those sense impressions are turned into reactions, conduct, knowledge, even memories. By whom, or what? The TV itself? The room? The impressions themselves, reaching some kind of mysterious critical mass in some way we can't quite explain, and so we discard the whole process, even as we rely on it? Objectively, Hume is right. But as Kierkegaard pointed out, objectively is not the way we examine our own existence. We cannot stand outside of our own existence an examine it as we would a vase or a favorite pet. So, while objectively we may say we cannot observe the activities of this thing we call a self, that is because objectively, we are speaking nonsense.
Kierkegaard sardonically wrote of the man who got so outside his own existence, determined to be objective about observing all things, that he awoke one morning to find he no longer existed! It was his way of pointing out that the truth which is most important to us has less to do with Aristotelian observation of data, and more to do with the Socratic concern "How should we then live?" (This is where Tillich grounds his idea of God's relationship to humanity being one of "ultimate concern" to the individual. What, asks Tillich, could be more important to one's existence, than the relationship with God?) This emphasis on the individual seems very natural to us, but it is indeed very much the product of Romaniticism. But that is still another essay!
The issues of the difficulty of our observation raised by Hume are left intact here. The answer of Kant is left intact as well. What shifts is the ground of the discussion. What is most important to the individual: establishing the validity of our construct of consciousness (which both Hume and Dennett, after him, were concerned with), or establishing the validity of our existence in the world? What is the truth of your love for your family? Can you establish it scientifically, empirically? Can you establish it for me so that I can objectively see the validity of your love? (Does that even matter to you?) And yet which is more important to you as a matter of "ultimate concern"?
So it comes down, finally, to a matter of language, where Hume almost 200 years ago. What is truth? Well, what do you mean by the term?
what, in Wittgenstein's terms, language game are you playing? He explains the matter this way:
In religion every level of devoutness must have its appropriate form of expression, which has no sense at a lower level. This doctrine, which means something at a higher level, is null and void for someone who is still at the lower level; he can only understand it wrongly and so these words are not valid for such a person.Empiricists like Dennett are still struggling with Hume (a far better philosopher, W.V.O. Quine, for example, worked to save empiricism from Hume's nihilism, without resorting to Kant's idealist escape hatch). Materialists like Dennett insist truth can be known only in terms of the material universe, but Hume leaves them unable even to talk about that universe, without resorting in some way to Kant's idealism.
Which leaves us in a rather dry place. Wittgenstein, ironically, can point a way out, by reference to that "subjectivity of truth" we started with:
Christianity is not a doctrine, not, I mean, a theory about what has happened and will happen to the human soul, but a description of something that actually takes place in human life. For 'consciousness of sin' is a real event and so are despair and salvation through faith. Those who speak of such things (Bunyan for instance) are simply describing what has happened to them whatever gloss anyone may want to put on it.