Matthew Southey half-baked half-thoughts    About    Archive    Feed

Science/Religion: Apophasis and Cataphasis

What follows is an attempt to compare scientific discourse with descriptions of mystical experiences by analyzing the information contained in statements. The categories apophasis and cataphasis come from religious studies but they are not parochial. I especially do not mean to confuse attributes (anything like “quantum physics is mysterious and consciousness is mysterious, therefore they must be related”). This is a discussion about the information available in rhetorical descriptions – statements differ in specificity and therefore give more or less shape to the objects being described. How statements, predictions, hypotheses and other linguistic statements attempt to model the world is fascinating – the philosophy of science has a lot to say about the accuracy of such statements.


In religion, there are two ways of talking about the divine: through negative statements (e.g. “god does not have a shape”), or positive statements (e.g. “god is love”). The first type of statement is apophatic (lat. via negativa), or description through negation. The second type of statement is cataphatic (lat. via positiva), or description through positive statements. Both of these methods can be found throughout the world religions.

Historically, cataphasis and apophasis have been used torward different theological ends. Apophatic statements, such as the famous Vedantic expression “neti neti” (“not this, not this”), are often used to convey the ineffability of an experience. The neoplatonist Proclus, suggested that cataphatic statements direct us toward the absolute, while apophatic statements are actually closer to the absolute. Advaita Vedantins often assert that the godhead (skt. nirguna brahman) is beyond all attributes and qualities. According to such philosophers, we will never be able to describe the ultimate with conventional language because they are categorically different. Using cataphatic language is seen as reducing mystical experiences to the mundane. Despite the categorical difference, some claim that words act as pointers while acknowledging that such pointers are insufficient in themselves. This is the classic “finger pointing at the moon”, where the finger is simply a symbol that directs one towards the ineffable.

The only reason to use language is to delineate the experience of the ultimate, and if the ultimate is beyond any and all words, and words do not even point, then why speak? If words are completely useless there is no reason to say anything at all. As the last proposition of Wittgenstein’s Tractatus says, “Whereof one cannot speak, thereof one must be silent”.


It is always easier to say something apophatic about an object. That is, there are more ways in which something is not than ways in which something is. Say there is a random number; you are more likely to be correct in saying that the random number is not 78429373 than it is 78429373. Out of all the numbers, there is only one correct answer and a near-infinity of wrong ones. Regardless of how many correct answers there can be to a question (there are often many), there will still be more wrong ones. Apophatic and deconstructive literature take advantage of this asymmetry by always offering negative statements. This deconstruction can serve a useful purpose when a truth-claim is offered without sufficient evidence. Nevertheless, answers do exist, and it is important to make reasoned, rigorous attempts at hitting them. Cataphasis is a risky endeavor, but the alternative is repeating “neti neti” until the heat death of the universe.

For some, the divine can be described through language but it is extremely difficult to do so. In the “random number” metaphor, the ability to hit the right sequence of digits is near impossible but a correctn number does exist. This stance can safely harbor both apophatic and cataphatic statements. This stance is similar to modern forms of rationality and science.

The philosopher of science Karl Popper says that science converges upon truth but never lays claim to it. That is, we can know what is not true but we cannot know if something is fundamentally true. This relates to Popper’s idea of falsification: we are not looking for positive support that a theory is correct,we are looking for a theory that withstands sustained attempts to be proven wrong. A good theory must be falsifiable (capable of being wrong), and then it must withstand attempts to be shown incorrect. Any theory is provisional upon not having yet been falsified. In the ideal scenario, specific hypotheses are offered and then rigorously evaluated. Either the hypothesis survives and becomes a provisional “truth” or it becomes knowledge of what is not true. Einstein is paraphrased as saying “no amount of experimentation can ever prove me right; a single experiment can prove me wrong”.

Scientific truths are similar to a number with infinite digits (like pi): we can only ever possess a limited understanding, and yet accuracy increases. Science converges upon truth but can never definitively hold it. Just like new digits of pi, it is necessary that future theories capture and explain the observational data of the past (the first digits of pi will always be 3.14 regardless of new digits discovered). For example, classical physics is fully explainable within quantum physics; but the reverse isn’t true, classical physics does not predict or explain quantum behavior. However, calculating the trajectory of an artillery shell by the interaction of subatomic particles is computationally infeasible. Classical physics is useful for calculating the behavior of large objects quickly and at a highly reduced computational expense. Some physicists consider quantum physics to be “more real” than classical physics, since quantum physics contains and extends upon previous theories. For more on this see Yudkowsky’s essay where he discusses how our maps are multi-level while the territory is single-level. This single-level is always converged upon but never reached).


The asymmetry between apophatic and cataphatic statements is reflected in the counter-intuitive claim that “the least probable theory is preferred to the most probable theory”. The least probable theory is the one with the highest information content and the one most open to falsification. Therefore, truisms like “all swans have a color” are less preferable than a statement that is more informative (and thereby has more ways of being wrong) like “all swans are white”. The least probable theory that withstands attempts to falsify it is the theory that becomes a provisional truth.

It is only the possible negation of cataphatic statements or the possible affirmation of apophatic statements that are scientific. A cataphatic hypothesis that “vaccines cause alzheimers” is scientific because it is possible to prove that vaccines do not cause alzheimers. An apophatic hypothesis that “vaccines do not cause alzheimers” is scientific because it is possible to prove that vaccines do cause alzheimers. Scientific statement must have the ability to be readily falsified in order to be scientific. However, there is an obvious asymmetry in the results of any tested statement: cataphatic answers are worth more than apophatic answers. There is more information in “vaccines cause alzheimers” than in “vaccines don’t cause alzheimers”. Alzheimers is caused by a limited number of things. Let’s say there are 10 causes of alzheimers – by getting a positive answer we know the epistemic status of 1/10 possibilities. On the other hand, if we learn that “vaccines don’t cause alzheimers” we are learning the epistemic status of 1/(near-infinity) hypotheses. Apophatic knowledge is still useful, but not nearly as useful as cataphatic knowledge.

Of course, possible hypotheses are not like the number line, there is not really a near-infinite ways of being wrong – this is hyperbole. Humans are pretty good at narrowing down possible hypotheses based on evidence. Formalized methods of weighing evidence exist, such as bayesian methods, but for day-to-day problems humans naturally narrow possibilities. Even if you are a great bayesian-rationalist, there will still be more possible ways of being wrong – this is what makes most scientific questions difficult and non-trivial (even for geniuses). It is possible to imagine that a future superintelligence will develop near-perfect bayesian weighing of evidence so that the number of possible hypotheses is drastically reduced. Testing of the hypotheses would be greatly reduced, and the pace of science would greatly increase. In this possible future, for this particular superintelligence, there may be more right hypotheses than there are wrong hypotheses! Scientific problems would be easy for this being. In this essay, I am only talking about our limited and feeble human capabilities of reason.


In religion, the cataphatic statement that brahman (the ultimate) is truth, consciousness, and bliss (skt. satcitananda) epistemically defines mystical experiences as “true” or “false”. If you take this statement as accurate, and have a painful experience where you are beset by delusions of shape-shifting demons, you can be quite certain that what you experienced was not brahman. If your experience was characterized by satcitananda then you could be fairly confident you were on the right track.

If brahman was apophatically defined as “not gendered”, this statement could only define (give information about) your experience if your experience involved gender – and even then it could only deny your experience as false, it has no confirming power. There is more information in cataphatic descriptions of the ultimate, there are more ways in which one’s experience can match (or fail to match)a statement. This is true if one believe’s that there there is some connection between mystical experience and language. If there isn’t any connection, why speak? And if language serves as pointers to the absolute – is there a way to analyze the information and efficacy of pointer-statements?

What is fundamentally unusually about religious experience, is that it is difficult to “objectively” access the referent of the mystical experience. It either seems to be completely subjective, or at least intimately connected to subjectivity (a la Jeff Kripal). Either way, there is not a common object that can be reliably tested through replicated studies. Correlates of the experience (physiological/brain states) can inform analysis but the experience itself is still only available through self-reporting (and therefore language). Looking closely at the information content of statements is important for it gives the varieties of experience outline and shape.

Next: Phenomena & noumena, intension & extension, exploded & imploded selves, logic, and the guidance of experience