# Derivation vs. proof

I imagine that Ramanujan and many other gifted mathematicians achieved results by derivation, meaning they followed a trail that led to a result, then proved it later if at all. Is this so and if so, why aren't derivations emphasized in education instead of proofs? I find the statement of a theorem without context and without relevance (bearing, concern, objective) and without origin, to be incomprehensible. Is it just me? Do folks who seem to absorb these like sponges have some special facility that I lack? I seem to recall reading that when Newton was criticized for not showing how he came to his conclusions, he said that when his building is finished, he takes down the scaffolding. Doesn't the derivation of a result reveal its bearing? I rarely find that a proof does.

Thank you,

-
Welcome to MSE! Some users here find ALL CAPS speak to be a bit rude, so I recommend editing them out of your question. If you feel they should be emphasized, perhaps use this or this. Also, your name is at the bottom of every post, so it's not necessary for you to sign the end of the post. – mixedmath May 5 '12 at 4:34
But they are! Research lectures have derivations, in your sense of the term, not proofs. Much of the content of a good mathematical lecture consists of derivations. Sadly, because of a fashion for conciseness, published papers are more proof-like. – André Nicolas May 5 '12 at 4:38
@George It would help to give some specific examples. – Bill Dubuque May 5 '12 at 4:51
Euler should be listed along with Ramanujan in this one. ${}\qquad{}$ – Michael Hardy Nov 27 '15 at 19:09

I know this is an ancient question, but given that there are no answers, let me give it a shot. As a disclaimer, this is a rather philosophical (or soft, if you may) answer, which I believe is justified by that the question is a soft one.

First things first, I believe calling "following a trail that led to a result" a "derivation" is misleading: the word "derivation" somewhat has a connotation of a construction, or taking explicit steps to find some other thing from one thing. So instead let us call what you call a "derivation" "intuition". Of course a better way to go about this is not to use such problematic words and make use of a spectrum that includes philosophy, mathematics and physics (and quite possibly other disciplines as well, but this whole idea is all so cumbersome to write down for now) (In your case I would consider you closer to physics, for instance, whereas I am more towards philosophy, and I am considering the spectrum to be somewhat of linear nature, which is unjustified). Indeed as you have said, oftentimes intuition comes before concrete proofs in mathematics: one makes sure that something has to be the case, and then tries to prove it. This is more so in the frontiers of mathematical knowledge, for there are no nice roads to walk on there; only muddy (and often dark) trails.

Next, we should determine what kind of education you are talking about. For your concern to make sense, I believe we should restrict education to education given by an institution: colleges, highschools etc.. So we disregard all self-studies, where one is basically by himself. For a similar reason, also disregard post-qualification exams in graduate life. With this restriction I can now talk about my understanding of the subject.

I have taken courses from mathematicians with diverse understandings of how a course should be thought, e.g. one doing every single proof and throwing in historical context every once in a while, another giving no proofs but only sketches, or ideas of proofs at most, another mostly demonstrating theorems on examples etc.. But in any of my studies I have seen that intuition and rigor go hand in hand, and that's what makes mathematics both very strict and also very flexible: otherwise either everything is trivial, or "can be seen from the picture" (as it happens in geometry), or else you lose track of what kind of objects you are working with (as it happens in analysis). Of course these are extreme (and to a greater extent false, consequently) statements, but in a "sentimental level" they have some truth in them. As long as there is human interaction I believe intuition is never left behind in communicating mathematics, though if you provide some clarification maybe I can modify my answer.

That leaves one final thing to be accounted, namely written mathematics. Again, from my experiences, mathematics addressed to a general audience, or expository mathematics usually does emphasize intuition, so that leaves mathematics textbooks for the [candidate] mathematicians, where everything has the potential to seem like cryptic and incomprehensible: out of the blue results, without any motivation provided for the reader that it is a crucial thing to learn about them. But here again, the answer lies in realizing that written mathematics expects a lot from its reader, until the reader has internalized what has been written, that is. Oftentimes, perhaps possibly because of the concise character of a mathematical script motivation, context and intuition is packed in one footnote; or such a part may lack altogether. In any case, mathematical writing expects you to develop your own intuition (and see that the results, when ordered that particular way (which may not be unique), makes a very nice trail to follow (even though muddy and dark it may be)). A quote from Gilbert Ryle, from his book The Concept of Mind, summarizes my point quite nicely I believe:

"Euclid's Elements are neither a sealed, nor an open, book to the schoolboy." (p. 56, the University of Chicago Press)

EDIT: I have just encountered a very explicit emphasis on intuition in Spivak's A Comprehensive Introduction to Differential Geometry, Vol. 1. On p. 84 he says:

"The tangent bundle is the true beginning of the study of differentiable manifolds, and you should not read further until you grok it.*"

"[...] [The word "grok"s] sense is nicely conveyed in The American Heritage Dictionary: "To understand profoundly through intuition or empathy"."

-
The definition of 'limit' used in calculus is wrong. Berkeley was arguing about nothing - the infinitesimals are neglected because they represent quantities that cannot be detected or measured. There's no contradiction. – mistermarko Nov 27 '15 at 19:42
How can a definition be wrong? That which is defined is defined to be so and so by the definition, no? Unfortunately I still did not read Berkeley's critique (to be honest it's not in my reading list for the next couple of months), so I cannot agree or disagree with you on that comment. – A. Alp Uzman Nov 28 '15 at 18:45

I can say that it is fairly true that when people (like Ramanujan himself), does not have a proper scholar background, might have a tendency to obtain a result, let us say "empirically". Even people like Carl Frederick Gauss, with proper scholar formation, got some of his most famous results that way (I am thinking on the Prime Number Theorem, as mentioned, for example, in the book "Prime Obsession", by John Derbyshire). However, as the previous answer hints about, if someone has obtained a result in mathematics, you want to know if it is actually true, or just the consequence of the particular path of derivation that was made. Come to think of it, at the age of 15, I derived algebraic expressions for the sum of the first "n" even numbers, and another for the first odd numbers, and the only way I had to know those expressions were correct, was by plugging in some integers on them, some very large, and with a pocket calculator as a backup. But the only way one can be really convinced such statements can be useful in general, came much later, when I learned mathematical induction (proof by induction).

Having said all that, in most classrooms a motivation for a certain mathematical subject, or particular theme is almost certainly exposed. This motivation may include what you call derivations, like in volumes obtained by triple integrals, where quite often a drawing is made of the differential volume elements, and it is mentioned the Riemann sum (as in the book from Earl. W. Swokowski "calculus with analytic geometry). Even Isaac Newton, who claimed, as you pointed out "that he takes down the scaffolding", if you read his book Philosophia Naturalis Principia Mathematica, and even if you are reading the original in latin, you can practically follow many results, derived quite often from geometric constructions; for example, section IV, Lemma III "De corporum circulari motu in mediis resistentibus" (which I think could be translated into "about bodies with circular movement in "resisting" -dragging- media), and then Newton draws a spiral and derives some geometrical consequences. Also his propagation of waves in fluids through a hole, is very intuitive, in "Der motu per fluida propagato", section VIII. And, of course, the parts that suggest what would become differential and integral calculus; for example, section II, "De motu corporum quibus resistitur in duplicata ratione velocitatum", where it is discussed that when a body moves in a resisting medium (and the resistance is the same all along), the velocity diminishes uniformly (see the picture; if it cannot be seen, the graphic contains a section of a hyperbola, with the "$x$, $y$" - ($t$, $v(t)$) axes, and this hyperbola is divided with vertical segments equally spaced in the "$x$" sense, to illustrate how a velocity decreases proportionally equal to equally elapsed $\Delta t$, at geometric points which Newton labeled B, k, l, m in the hyperbola and A, K, L, M over the "x" or time axis). Newton's expositions are more what you call "derivations" that actual proofs in our modern sense. It is also clear from Isaac Newton's style of writing, that this work has more "flavor" (so to speak) to philosophy than to what we would call modern mathematics, but that is of course our historic, hindsight perspective.

In conclusion, it depends on the particular teacher or researcher, but I don't think in modern times intuition has been abandoned, and contrariwise used for the sake of exposition, but also, on the other hand, scientific rigor is not forsaken either, and the tendency is to mix them both every time more and more.

-