How to Recognize a Sound Philosophy

In mathematics, a new theorem can be subjected to a rigorous proof, which any mathematician in the field can verify. In science, a hypothesis can be shown to be in agreement with data to some desired level of probability. In philosophy, things are not quite so rigorous. But that does not mean one philosophical argument is as good as any other! A philosophy can be fully solid and sound, or it can be complete nonsense – or, like most things in life, it can be somewhere in between. Fortunately, there are reliable ways to judge, through careful analysis, the soundness of a philosophical position.

General Methods

To tell when a philosophical theory is sound, first it must be stated carefully, using terminology that we can all agree on. We’re going to subject the theory to a variety of tests, and we can’t do that reliably if we’re not sure exactly what the theory is. In fact, that’s the first thing to look for:

Clarity: a sound philosophical theory can be stated clearly, so that anyone with a reasonable background can understand what it means. Ideally it can be encapsulated in a single sentence, even if some of the terms bear some further elaboration, as we did in the Introduction.

A theory stated clearly like this will hold still while you examine it; it won’t squirm and morph into different forms to pass first this test and then that test. That seems like an obvious requirement, and yet particularly in casual discussions, people do this all the time; they have some vague, half-developed theory of identity in mind, and when challenged with some test, they simply modify it to pass that one test, even if so doing invalidates earlier results (which are conveniently forgotten).

So, the first test for any philosophical theory is that it can be stated clearly and unambiguously. Once you have that, there are a number of other features it should have. Let’s consider each of these in turn.

Common Sense Consistency: as noted in the last essay, a good philosophical theory should agree with common sense in common situations with which we have robust experience. For example, in the area of personal identity, we would expect all of the following every-day observations to hold:

• I am (almost entirely) the same person from moment to moment.
• I am (almost entirely) the same person after a nap, a good night’s sleep, or a successful surgical operation.
• Who a person is can change somewhat over time, or after some formative event.
• Twin siblings are two different people.

As pointed out previously, we must take great care not to include tests about which we have no experience, such as “a person exists in exactly one body” or “a person can’t be in two places at once.” These statements are actually conclusions we might be able to draw from a theory, and not something we should assume going in; they may well reflect limitations of our current technology, rather than any deep truths about the nature of identity itself. We can tell the difference by asking: is there any conceivable technology or magic which might make this possible? For example, could we imagine advanced aliens where one person occupies six bodies, or a time traveler who can go back in time and meet herself? If such things are even imaginable, but outside the realm of our current experience, then these are not situations where common sense can apply.

Logical Consistency: in logic and mathematics, an identity relation is expected to have several key properties. These include:

• Reflexivity: any entity is identical to itself. Expressed in symbols, if x refers to some entity, then x = x.

• Symmetry: if x is identical to y, then y is identical to x. In symbols, x = y implies y = x.

• Transitivity: If x is identical to y and y is identical to z, then x is identical to z (x = y and y = z implies x = z).

• Indiscernibility: If two entities are identical, then they share all the same properties; they are indistinguishable in every respect.

Of course philosophy is not mathematics, and these properties may not be absolute requirements. But the more of these properties are violated by a particular theory, the more we should suspect that what that theory provides is not an identity relation.

Moreover, it’s worth noting that these properties were all based originally on Boolean logic; but they are easily extended to modern fuzzy logic. For example, the Symmetry relation should really be stated “to the extent that x is identical to y, y is identical to x.” This use of fuzzy logic may also save the Indiscernibility relation, which as stated above is much too strong to be of any practical use in personal identity. But in its fuzzy form, “two entities are identical to the extent that all their properties are the same,” it may still be applicable, as it would allow one to be mostly the same person from day to day, despite some minor differences.

Internal Consistency: a sound philosophy should be internally consistent and coherent; it should not lead to self-contradictions, or require assuming what it claims to prove (i.e. rely on circular reasoning).

Empirical Consistency: a sound philosophy, to the extent that it implies anything about the physical world, should be consistent with empirical data. In other words, if a philosophy intersects with science, it should be consistent with current scientific understanding. For example, one might base a theory of identity on the idea that there is an immaterial soul that connects to and controls the body through the pineal gland in middle of the brain; but this is not consistent with our scientific understanding of how the mind and brain work. Such a theory would lack empirical consistency.

Comprehensiveness: a good philosophical theory is able to address all aspects of the topic of interest. In the case of personal identity, it should be able to answer “are entities x and y the same person?” for any scenario we can dream up, rather than only providing answers in some cases, and throwing up its hands in others.

How Does the Simple Solution Measure Up?

The tests above can be applied to any philosophical theory, including theories of identity. In a later essay, we’ll do exactly that to a variety of popular theories, such as “bodily identity” and even modern “branching identity” theory. But the main topic of this site is one particular theory in particular, the “simple solution” — though for a less opinionated label, we might call it “fuzzy information-content identity” theory. So let’s apply the above tests to this theory — both to illustrate how that is done, and see how well this simple theory holds up so far.

Clarity: fuzzy information-content identity can be stated as follows:

Two things share the same identity to the extent that their relevant information content is the same.

We already gave this statement, and elaborated on some of the terminology, in the Introduction. The statement is short and clear; it leaves little wiggle-room for shifting interpretations. So far so good.

Common Sense Consistency: we could have an entire section on this, and probably should. But for now let’s just try out a handful scenarios as given earlier in this essay, and see how our theory applies:

• I am (almost entirely) the same person from moment to moment because the information content of my mind (my knowledge, personality, memories, psychological traits, etc.) is almost entirely the same, changed in only minute ways by moment-to-moment experience.
• I am (almost entirely) the same person after a nap, a good night’s sleep, or a successful surgical operation because, again, the content of my mind is virtually unchanged by these events.
• Who a person is can change somewhat over time, or after some formative event because the content of their minds changes somewhat with time and experiences.
• Twin siblings are two different people because, while they may look very similar and be genetically identical, the content of their minds are completely different; they have different memories, personalities, knowledge, hopes and dreams, etc.

So, in common situations about which we all have real experience, the simple solution holds up just fine. (Not all theories of personal identity do, as we’ll see later!)

Logical Consistency: it’s easy to see that the way we have defined identity here, based on the information content, all the expected logical properties hold:

• Reflexivity: x = x because the information content of x is, obviously, exactly the same as the information content of x.

• Symmetry: to the extent that x = y, i.e. the information content of x matches the information content of y, then the converse is also true.

• Transitivity: this one is more interesting, but it holds by the rules of modern fuzzy logic. To the extent that x‘s information content matches y‘s, and y‘s matches z‘s, then x will match z‘s (in the amount of x = y AND y = z, or better). If we were to put numbers on it: if x and y have a 95% match, and y and z have a 95% match, then the match between x and z will be at least 95% * 95% = 90.25%.

• Indiscernibility: because of the word “relevant” in our definition of identity, we have to be somewhat careful with this one; we would say that two entities are indistinguishable only in relevant aspects. For example, as stories, my copy of Moby Dick and your copy of Moby Dick are functionally indiscernible provided the text is the same, even if I can distinguish them in other ways (such as mine having my name scrawled in the front cover). If people could be backed up and restored, we would say that two incarnations of the same backup are (at the moment they are switched on) indiscernible, even if one is sitting in chair A while the other sits in chair B.

Internal Consistency: this one is harder to test, because any inconsistencies are likely to be unique to each theory under consideration. However, the current theory has been advocated for nearly two decades, and no internal inconsistencies have been found yet.

Empirical Consistency: fuzzy information-content theory is in no way incompatible with scientific understanding. Applied to personal identity, in particular, it assumes materialism, i.e., that the mind is a product of the brain, and all the information in our psychological makeup is encoded somehow in the brain (and possibly some parts of the body, e.g., the peripheral nervous system or endocrine system). This materialistic view is strongly supported by neuroscience. Applied to inanimate objects, it will often come down to observing that objects are made of atoms, and atoms of a particular element and isotope are indistinguishable; there is no immaterial “soul” of a something like a hammer. This is again completely consistent with modern physics.

Comprehensiveness: the theory can answer “are entities x and y the same person?” for any scenario we can dream up, simply by looking at the extent to which the information content of the minds of x and y are the same. In fact, it can go further, and apply to other identity questions too, as long as we can define what the “relevant information content” is. (We’ll dedicate a whole essay on this later, exploring things like the famous Ship of Theseus — a hypothetical ship which is the very ship used by Theseus himself, though every component on it has been replaced at least once.)

These tests for a sound philosophy fit the Simple Solution so well, you may be forgiven for suspecting that they were chosen to support it. But in fact, the exact opposite is true – the solution was developed specifically to pass all these tests, as any good theory should be. Whatever the true nature of the world, we should be able to explain it in a way that is clear, logical, and consistent. If fuzzy information-content identity does not yet resonate with you, I urge you to apply the above tests to your own ideas, and see where they they may need revision.