Friday, December 22, 2006

Just What Is the Principle of the Development of Doctrine, Anyway?

In a singularly unhelpful comment on the running dialogue regarding the Principle of the Development of Doctrine (PDD), one reader suggested that, if I am right about development being Non-ampliative, then we might as well replace the Magisterium with a computer. To say such a thing is to simultaneously misunderstand the nature of deductive reasoning and the nature of the Magisterium (not to mention the nature of computers, but that is another tale). In thinking about possible ways to reply to such a thing, however, I set myself to thinking about the central focus of these posts: the PDD itself. It occurred to me that some readers have a rather vague notion of what that Principle is in the first place.

Let me start things off here with a quotation from Fr. Richard John Neuhaus, a particular theological hero of mine, summarizing the central points of Blessed John Henry Newman's views regarding the development of doctrine. It is worth pointing out that Dr. Mike Liccione also refers to this passage in his own extraordinarily helpful discussion of the development of doctrine at Pontifications. Fr. Neuhaus writes:
Recall Cardinal Newman’s reflection on the development of doctrine, a reflection that has been incorporated by magisterial teaching. He suggested seven marks of authentic development: authentic development preserves the Church’s apostolic form; it reflects continuity of principles in testing the unknown by the known; it demonstrates the power to assimilate what is true, even in what is posited against it; it follows a logical sequence; it anticipates future developments; it conserves past developments; and, throughout, it claims and demonstrates the vigor of teaching authority. And thus it is, said St. Vincent of Lerins in the fifth century, that in authentic development of doctrine nothing presents itself in the Church’s old age that was not latent in her youth. Such was the truth discovered by Augustine, a truth "ever ancient, ever new."
While it is nice to see the expression "follows a logical sequence" in there, one must of course admit that inductive reasoning is "logical" in precisely the same sense that deductive reasoning is, but I would also draw attention to the thought of St. Vincent that authentic development of doctrine adds nothing to what was already latent in the youth of the Church. This is an explicitly non-ampliative understanding of "authentic development of doctrine", and only deductive inference is non-ampliative in this way. This is the understanding of "authentic development of doctrine" both of a fifth-century writer (and, arguably, of Saint Augustine) and of Newman himself.

Having said this, however, I am not at all sure that there is not sufficient common ground between my friend Mike Liccione and myself to reach some sort of a consensus on this issue. Just for starters, it was his posting on the development of doctrine at Pontifications that helped me to form my own views, or at least to give them a more substantial form than they had had before. It occurs to me that the principal difference between us, at least as the two of us appear to be construing it in our own posts, has to do with what might be describable as something like "cognitive leaps" in the developmental process. This was something that I myself had hinted at in my previous post when I talked about the differences between inference and interpretation. I think some of the commenters (at least at this blog) may have had something like this in mind, too, when they mentioned such teachings as the Assumption and the Immaculate Conception as exceptions to my rule that doctrinal development be Non-ampliative. In my view, the teachings on the Assumption and on the Immaculate Conception do not, in fact, represent examples of doctrinal development anyway--they are, rather straightforward teachings of the Church--but that is actually not a crucial issue in this debate (at least, not yet).

If I am right that there exists some common ground here, I think one way to highlight what I think that common ground might be would be to turn the discussion for a moment to Aristotle's Analytics. The Analytics are traditionally divided into two works (though this division is not Aristotle's, and he himself always refers only to a single work under the title Analutika), the Prior Analytics and the Posterior Analytics. In the Prior Analytics Aristotle lays out in typically exacting detail his theory of syllogistic logic. Aristotle was particularly proud of his accomplishment in the Prior Analytics: he was the first philosopher to construct a formal theory of inference, and it was--and is--a perfectly good formalization. Indeed, it was virtually the only formal logic in use until relatively recently, and it continues to be taught in logic classes. It has been augmented in some ways by the work of modern formal logicians, but it has not been superseded in any significant sense. The Posterior Analytics is more difficult to classify. Some say it is a treatise in epistemology, others speak of it almost as though it were a work in the philosophy of science. Aristotle was particularly interested in scientific topics, and so there may not have been a very sharp distinction in his mind between those two areas in the first place. (It is somewhat anachronistic to speak of "philosophy of science" in Aristotle's case, however, so one must be willing to blur some distinctions anyway when talking about these things.) The relationship between the Prior and Posterior Analytics is controversial among scholars of Aristotle. Aristotle was himself a very active scientist, in the sense of "natural philosopher" if not in our own sense of the term, and his work in the foundations of biology, in particular, was seminal, influencing the development of that science right up to the time of Darwin himself, who found Aristotle's taxonomy of species superior to Linnaeus'. Scholars have noticed, however, that Aristotle does not employ, in these biological treatises, any of the principles of inference that he lays down in the Prior Analytics, nor does he employ the syllogism in any of his other scientific writings. This has prompted some to suggest that his own scientific work was more along the lines of foundational speculation and data gathering, and some scholars have suggested that the principals he follows in his own scientific work are those established in the Posterior Analaytics, where the subject matter has more to do with the establishing of axioms, definitions, and scientific procedures. One line of interpretation holds that Aristotle saw science as a three-step process: first we establish axioms and definitions, then we gather data, then we organize the date in accordance with rational principles. Only the final stage, the organization of the data (perhaps for pedagogical purposes), would employ the principles of the syllogism. Everything else would be governed by the inference pattern that Aristotle calls epagogĂȘ and that we call induction.

Consider the following two inferences.
1. "Mammal" is necessarily true of all humans.
2. "Animal" is necessarily true of all mammals.
3. Hence, "Animal" is necessarily true of all humans.
Constrast this inference with this one:
1. Socrates is a human, and he is also an animal.
2. Plato is a human, and he is also an animal.
3. Callicles is a human, and he is also an animal.
4. Coriscus is a human, and he is also an animal.
5. Hence, anything that is a human is also an animal.
The first inference is an example of a syllogism in the Aristotelian (categorical) style. The second is an induction. Both arguments give us rational warrant for asserting that all humans are animals, but only the first argument actually proves, with logical necessity, that all humans are animals. The difference lies in what Aristotle called the meson, and that later logicians called the "middle term". Aristotle held that the meson served a special, explanatory role in inference: it gives us the reason why something is true. The so-called "middle term" (here, "mammal") represents a class of things that connects the "major term" (here, "animal") to the "minor term (here, "human") by a kind of overlapping relation: some things that are animals are also mammals, but not all of them; however, all mammals are necessarily also animals. This relation guarantees that it is quite impossible for something to be a human being and not be an animal.

Now every deduction, if it is to be sound, must meet two conditions. It must be valid in form, and its premises must be true. Putting validity aside for the moment, how are we to know that the premises are, indeed, true? One way is to provide a proof of each premise. There are two ways we can give rational warrant for a premise: we may supply another deduction in which the premise we are trying to prove appears as the conclusion, or we may supply an induction that leads to the premise we are trying to support. Obviously only the first will count as a "proof" of the truth of the premise in question, but we cannot demand that deductions be given for every premise, because that will lead to an infinite procedure of premise-proving. We don't want to have an infinitely long sequence of deductions because these proofs, for Aristotle, are supposed to be explanatory, that is, they are supposed to give the reason why something is true. Anything that is infinite is necessarily incomplete, and so if we demand that all proofs be deductive then we can never have a complete explanation of anything. Some of our reasons for believing certain premises are going to have to be inductive.

For example, we may discover inductively that all human beings are mammals. It might be quite impossible for us to actually observe all the humans who exist right now, and it is obviously impossible for us to observe future generations of humans who don't even exist yet, and so to make the claim "All human beings are mammals" is obviously an induction, and yet we regard it as necessarily true because it represents a claim about the essential nature of human beings, and the essential nature of a thing is true of it necessarily. That means we can know it with certainty in spite of the fact that it is discoverable only by induction.

Now, this is a rather handy feature of Aristotle's essentialist metaphysics, and it is not available to contemporary scientists, but it is certainly available to Thomists and other essentialists in the Christian tradition. We agree with Aristotle that there are things that we can know with certainty on the basis of inductive reasoning, and I suspect that this is the sort of thing that Mike and others may have in mind when they argue that doctrine develops inductively. Or, to put it another way, if this is what they do have in mind, then I don't disagree with them, other than to say that, just because there are some inductions in there, we are not entitled to assume that the only sort of proof of the doctrine available is going to be inductive. As in the case of the proof that all humans are animals, both an inductive and a deductive inference may be available, and the deductive is always to be preferred, and for two reasons. First, it gives the reason why, hence proving its conclusion beyond any rational doubt; and second, the inductive inferential pattern can only be regarded as acceptable if it is grounded in necessary principles in the first place, and those will have to come from the teaching authority of the Magisterium, that is, they will have to be matters of interpretation, not inference.

Aristotle, in addition to being a metaphysical realist, was also a rationalist. That is, he was not an empiricist, which means that he did not think that we come to have knowledge of first principles via empirical induction, even though epagogĂȘ clearly plays a role in his story of how we come to have knowledge of first principles. But our cognitive grasp of the essences of things is a function of nous, our intellect, and it is entirely separate from the operation of perception. We grasp the universal that is present in the particular in a way that is explicitly rejected by contemporary empirical science. It should come as no surprise, then, that the Church should recognize a means of coming to have knowledge of the first principles of inferences by means other than mere empirical inductions made on scriptural texts. That would be tantamount to sola scriptura. On the contrary, it is the divine authority given to the Church herself that enables her to make pronouncements about doctrine that are infallibly (necessarily) true, and this is not a matter of inference, but of inspiration. There may be a process, similar to that described by Aristotle, in which we must lay out the results, as it were, of our research into these things that gives rise to an appearance of "cognitive leaps" in doctrine, but a full process of preparing the case will always enable us to put together a perfectly sound deductive proof that the teaching is true, and this process will always be non-ampliative.

Whether this represents real common ground between me and Mike will probably become clear only after further discussion, but I suspect that we are not far apart. I certainly trust in his intelligence and erudition to make clearer anything that I have said that may be worth holding on to and to help me to see my way clear to rejecting anything that I have said that is untenable.


zippy said...

I am willing to let my modest understanding of computers, deduction, and Magisterial authority (and my modest accomplishments in those areas) speak for themselves; and I am sorry you found my comments unhelpful. But if you aren't saying that once the initial body of doctrine was established at the time of the Apostles all additional development of doctrine must proceed deductively in order to in fact be doctrinal[**], then I really have no idea what you are saying. If all you are saying is that the development of doctrine must be logically consistent then I agree, but your requirements w.r.t. underdetermination seem to be saying more than that.

[**] Where "proceed deductively" does not mean "must be discovered as a matter of intellectual history only by deductive methods" but rather "must, once discovered, conform without underdetermination to a deductive formalism connected by other deductive formalisms to that initial body of doctrine". Or something like that.

(Unfortunately or perhaps fortunately, depending on one's disposition toward my participation in the discussion, my participation will necessarily be limited from here forward. But thanks for the stimulating discussion).

Scott Carson said...

Don't take it personally--I doubt that there are many folks out there who take what I find to be helpful as anything like a yardstick of what actually is helpful. Your comments are very much appreciated, and I hope that you will have opportunity to comment more! I suspect that we are not that far apart on this one, since I rarely find anything to object to at your blog and, indeed, I read it regularly for enlightenment and instruction.

zippy said...

Thank you Scott, and no worries. I look forward to returning to this subject (at least to thinking about this subject!) in the new year. And I agree that our differences, to the extent they are indeed substantive differences, are probably not large at all (and may well as a practical matter be irrelevant). (In a way that is part of what makes this particular discussion so interesting).

Mike L said...


I'll probably post my reply by tomorrow night, if not before then. For now, it seems we are making progress.

Merry Christmas,