Despite the richness that mathematics imparts to life, it remains a hated and difficult subject to many. I feel that the difficulty stems from the early and often permanent disconnect between math and reality. It is hard to memorize that the reciprocals of bigger numbers are smaller, while it is fun to figure out that if you had more people sharing a pizza, you get a smaller slice. Figuring out is fun, memorizing — not so much. Mathematics, being a formal representation of the patterns in reality, doesn’t put too much emphasis on the figuring out part, and it is plain lost on many. To repeat that statement with mathematical precision — math is syntactically rich and rigorous, but semantically weak. Syntax can build on itself, and often shake off its semantic riders like an unruly horse. Worse, it can metamorphose into different semantic forms that look vastly different from one another. It takes a student a few years to notice that complex numbers, vector algebra, coordinate geometry, linear algebra and trigonometry are all essentially different syntactical descriptions of Euclidean geometry. Those who excel in mathematics are, I presume, the ones who have developed their own semantic perspectives to rein in the seemingly wild syntactical beast.
Physics also can provide beautiful semantic contexts to the empty formalisms of advanced mathematics. Look at Minkowski space and Riemannian geometry, for instance, and how Einstein turned them into descriptions of our perceived reality. In addition to providing semantics to mathematical formalism, science also promotes a worldview based on critical thinking and a ferociously scrupulous scientific integrity. It is an attitude of examining one’s conclusions, assumptions and hypotheses mercilessly to convince oneself that nothing has been overlooked. Nowhere is this nitpicking obsession more evident than in experimental physics. Physicists report their measurements with two sets of errors — a statistical error representing the fact that they have made only a finite number of observations, and a systematic error that is supposed to account for the inaccuracies in methodology, assumptions etc.
We may find it interesting to look at the counterpart of this scientific integrity in our neck of the woods — quantitative finance, which decorates the syntactical edifice of stochastic calculus with dollar-and-cents semantics, of a kind that ends up in annual reports and generates performance bonuses. One might even say that it has a profound impact on the global economy as a whole. Given this impact, how do we assign errors and confidence levels to our results? To illustrate it with an example, when a trading system reports the P/L of a trade as, say, seven million, is it $7,000,000 +/- $5,000,000 or is it $7,000, 000 +/- $5000? The latter, clearly, holds more value for the financial institution and should be rewarded more than the former. We are aware of it. We estimate the errors in terms of the volatility and sensitivities of the returns and apply P/L reserves. But how do we handle other systematic errors? How do we measure the impact of our assumptions on market liquidity, information symmetry etc., and assign dollar values to the resulting errors? If we had been scrupulous about error propagations of this, perhaps the financial crisis of 2008 would not have come about.
Although mathematicians are, in general, free of such critical self-doubts as physicists — precisely because of a total disconnect between their syntactical wizardry and its semantic contexts, in my opinion — there are some who take the validity of their assumptions almost too seriously. I remember this professor of mine who taught us mathematical induction. After proving some minor theorem using it on the blackboard (yes it was before the era of whiteboards), he asked us whether he had proved it. We said, sure, he had done it right front of us. He then said, “Ah, but you should ask yourselves if mathematical induction is right.” If I think of him as a great mathematician, it is perhaps only because of the common romantic fancy of ours that glorifies our past teachers. But I am fairly certain that the recognition of the possible fallacy in my glorification is a direct result of the seeds he planted with his statement.
My professor may have taken this self-doubt business too far; it is perhaps not healthy or practical to question the very backdrop of our rationality and logic. What is more important is to ensure the sanity of the results we arrive at, employing the formidable syntactical machinery at our disposal. The only way to maintain an attitude of healthy self-doubt and the consequent sanity checks is to jealously guard the connection between the patterns of reality and the formalisms in mathematics. And that, in my opinion, would be the right way to develop a love for math as well.