Principles of Sufficient Reason for Scientific Inquiries

Why do I suggest reading this, and how is it useful?

As a philosophy lover myself (I will someday add a writing about why I love philosophy and why it is so valuable), one of my favorite books is The World as Will and Representation by Schopenhauer. To conquer this book, there were many challenges and prerequisites, including Kantian books (Critique of Pure Reason to Critique of Practical Reason), Western philosophical history, and some knowledge of the Upanishads and Indian philosophy. But the most important prerequisite that I enjoyed the most was his 1847 revised work, The Fourfold Root of the Principle of Sufficient Reason.

This work beautifully manifested and criticized how earlier philosophers like Descartes and Leibniz confused and abused “causality” and “reason of cognition,” which created much of the chaos in metaphysical debates and in understanding their nature.

Today, I want to talk about the “metaphysics” of asking questions itself, and within this framework, how we can ask more coherent, mutually exclusive yet collectively exhaustive questions that could benefit PhD students and researchers like me.

1. Causal probe (Becoming)
2. Epistemic probe (Knowing)
3. Structural probe (Being)
4. Behavioral probe (Willing)

Asking scientific inquiries requires understanding their limits and scope

Some say there are no dumb questions… but I think there may be some. For example, asking the exact same question the previous person asked, asking a totally unrelated question outside the scope, or asking questions just to show off one’s knowledge might not be very helpful.

Asking and choosing the right question is our responsibility as both scientific researchers and engineers. What problems should we solve with this limited amount of time? Why should we solve this? How does this cell’s mechanism work?…

A scientific paper is often about answering an asked question, so it is very important to choose the right question that is worth our time and effort, while also maximizing its effective benefit to others and contributing to our knowledge in seeking truth. Simply saying… we may not want to end up asking something like, “What would happen if this chair moved 5 cm and stayed there for 100 years? Let’s observe!” or something that has already been asked.

But picking the right and good questions first requires understanding:

“What am I trying to ask, and what does it mean?”

The challenges of understanding a question are often because it is:

1. too complicated,
2. overlapped with other questions, so I myself do not know what I am asking,
3. too broad or too narrow.

When we talk about asking questions (which I think is, if not the only reason, one of the main reasons why our intelligence exists and what it is for, as Kant suggested), it is about seeking sufficient reasons for the phenomena that we observe or experience.

Fourfold Root of Sufficient Reason

What is sufficient reason? For me, sufficient reason is about seeking the reason why something has to be in that state or for that reason rather than another one. It has often been treated as one of the fundamental premises of logic and metaphysics.

Schopenhauer criticized Descartes, Spinoza, Leibniz, and other rationalists for making a fatal mistake: they often overlapped and abused different kinds of reasons that were supposed to be applied for different purposes. For example, physical reasons, or causality, must be distinguished from logical inferences, or knowing, and so forth.

Therefore, it is very likely that many researchers nowadays may make the same mistake.

1. The Principle of Sufficient Reason - “Becoming”

Satz vom zureichenden Grunde des Werdens, principium rationis sufficientis fiendi

It is about the natural becoming of objects. Objects here mean everything within our experience that we observe, and the “appearance and disappearance” of those objects are mutually connected within the boundaries and flow of time and space. This means that whenever a new state or object appears, there must be something that affected it. The former we call the cause (Ursache), and the latter we call the effect (Wirkung). But we do not know where this “first principle” originally started.

This question is more like asking, “Why did this happen?” in the normal scientific sense. It asks about causal interactions, causal probes of mechanisms, why one thing caused A, how A caused B, how B caused C, and so forth. In my research field of neural engineering, it might be something like: let’s say we find that stimulation (U) drives neural population dynamics (Z), and therefore produces a behavior (B).

2. The Principle of Sufficient Reason - “Knowing”

Satz vom zureichenden Grunde des Erkennens, principium rationis sufficientis cognoscendi

These reasons are about the logical grounds that govern or lead to a judgment, discrimination, or determination.In Schopenhauer’s view, thinking in the narrow sense is not merely the existence of abstract concepts inside consciousness. Rather, thinking occurs when two or more judgments are combined or separated through the many determinations and transformations described by logic and the theory of judgment. These clearly thought and expressed relationships between concepts constitute judgments.

I think this is often the type of question most confused with causal questions in science. For example, if my question asks which latent dimension causally generated real behavior, which belongs to becoming, the question of “knowing” instead asks what the grounds or underlying evidence are for believing that this dimension is behavior-relevant.

Another example would be a recent fMRI paper published in 2025, Samira et al., “BOLD signal changes can oppose oxygen metabolism across the human cortex,” Nature Neuroscience. This kind of work challenges a question of “knowing”: whether the evidence or measurement we relied on actually justifies the belief we formed from it. For example, if decoding accuracy is an indicator of performance, why is that true? Why is cross-validation a good indicator of generalization? Something like this. This means that the high accuracy of a decoder model can provide strong evidence for the Grund of knowing, but it does not necessarily explain a causal mechanism.

As a quote attributed here to Schopenhauer:

“Grounds for believing a proposition are not the same thing as causes of the phenomenon described by that proposition.”

3. The Principle of Sufficient Reason - “Being”

Satz vom zureichenden Grunde des Seins, Principium rationis sufficientis essendi

All objects are defined within two orthogonal dimensions, time and space, and within these dimensions, all objects exist in relationships with other objects as parts of a whole. Therefore, all parts are constrained and mutually defined by the existence of other parts. In space, Schopenhauer refers to this as position (Lage), and in time, as succession (Folge). He thought that geometrical and mathematical knowledge, as well as the kinds of questions humans can ask about them, are governed by these constrained relationships.

One example may be asking about the reason of being. By elucidating the relationship with other objects, without necessarily asking about causality, one might ask: “What structural constraints make a state possible?” For example, what kind of structure or constraint makes such a phenomenon inevitable?

4. The Principle of Sufficient Reason - “Acting / Willing”

Satz vom zureichenden Grunde des Handelns, Principium rationis sufficientis essendi

The last one concerns motivation in the psychological domain. The fourth object of representation is the direct object of inner sense: the “subject of willing” (Subjekt des Wollens). This is an object for the knowing subject, but unlike ordinary external objects, it does not appear in space and appears only in time. It refers to the subject not as something that knows, but as something that desires and wills.

So this question is about motivation and intention, or in modern terms, context, because context cannot necessarily be found within the system itself, but may instead come from outside the system. This question may therefore be relevant when asking: What external conditions, contexts, goals, policies, or rewards govern the grounds of action?

So, in summary, using the fourfold root, we can ask questions using the same framework in neuroscience and AI, thanks to Schopenhauer:

Fourfold Root of Sufficient Reason for Scientific Inquiry

1. Causal probe - What makes it happen? (Becoming)

   perturbation / latent neural dynamics / system identification

2. Epistemic probe - How do we know? (Knowing)

   prediction / statistics / generalization

3. Structural probe - Why is it possible? (Being)

   manifold constraints / geometry and rules / connectivity

4. Behavioral probe - When and who is it for? (Acting / Willing)

   goal / context / intention

Further reading

• Schopenhauer, Arthur. On the fourfold root of the principle of sufficient reason. Open Court Publishing, 1974.