Wavefronts, Box Diagrams, and the Product Rule: A Discovery Approach John W. Dawson, Jr. The Two-Year College Mathematics Journal, Vol. 11, No. 2. (Mar., 1980), pp. 102-106. Stable URL: http://links.jstor.org/sici?sici=0049-4925%28198003%2911%3A2%3C102%3AWBDATP%3E2.0.CO%3B2-X The Two-Year College Mathematics Journal is currently published by Mathematical Association of America. Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/maa.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected]. http://www.jstor.org Wed Aug 22 11:12:14 2007 Wavefronts, Box Diagrams, and the Product Rule: A Discovery Approach John W. Dawson, Jr. John W. Dawson, Jr., received his S.B. in mathematics from M.I.T., and his Ph.D. from the University of Michigan. He joined the Penn State faculty in 1972, transferring to York as an assistant professor in 1975. He is a specialist in mathematical logic. Of the many differentiation rules encountered in calculus, the product rule is perhaps the first to shake students' naive faith in the "simplicity" of mathematics. It is easy to refute their expectation that (fg)' = f'g', but counterexamples often seem more to bewilder than enlighten. Moreover, the correct formula is not easy to guess, except for the special case when one factor is constant; and although many students accept the formula (cg)' = cg' as agreeing with their intuition, few realize that this special case itself contradicts the equation (fg)' = f'g'. Later, having learned the correct general formula, some students laboriously apply it to all products, ignoring the erstwhile special case, or failing to recognize it as such! Such observations suggest that few students develop any real understanding of the product rule. Rather, having no idea why the rule takes the form it does, most students simply memorize it, adding it to their growing stockpile of unrelated facts. Of course, at some point they have probably seen the product rule derived through algebraic manipulation of the defining difference quotient. The usual proof is concise and rigorous, and students will probably assent to the correctness of each step. Nevertheless it is a hollow demonstration, one that verifies rather than explains, for it conveys no idea of how the formula, or the proof itself, might ever have been discovered. In particular, the rationale for transforming f ( x + Ax)g(x Ax) - f ( x ) g ( x ) to [ f ( x + Ax) - f ( x ) ] g ( x Ax) f ( x ) [ g ( x Ax) - g ( x ) ] via the simultaneous addition and subtraction of the term f ( x ) g ( x Ax) is mystifying to students; too often the step is justified merely as a clever trick that works. In short, the product rule provides a classic example of how a logically rigorous "proof" may fail to be psychologically convincing or enlightening. There is, however, a more heuristic approach to the product rule that leads to discovery of both the formula and its proof. This approach invokes the geometric representation of products as areas, an idea later to prove so important in integral calculus, while stressing certain physical considerations important in their own right. At the same time the mystery surrounding the algebraic manipulations is dispelled. Indeed, the only mystery left unexplained is the conspicuous absence of such an approach from most calculus texts. + + + + + The very word "calculus" provides a fitting point of departure: Latin for "pebble," it alludes to the counting stones used in early abacus-like devices, but it also calls to mind the ripple created by a pebble cast into water, and so suggests the familiar problem of computing the rate of increase of the area enclosed by a radially expanding circular wavefront. Although this problem appears in nearly every calculus text, it is usually considered only after introduction of the chain rule, and all too often it is presented as a routine exercise devoid of any particular significance. However, merely by factoring a difference of two squares, it is easy to show directly that and far from being surprising, the result is readily conjectured, since the area inside the ripple increases through the engulfing of space by the expanding wavefront, which at any given instant has length 27rr and is moving outward at speed d r / d t (Figure 0). Similarly, a radially expanding spherical membrane engulfs 3-dimensional space at a rate equal to the product of its instantaneous surface area and its instantaneous velocity. The physical intuitions involved are quite natural and general*, and students might well pass on to consider the rate at which area is engulfed by an "arbitrary" radially-expanding simple closed curve. It is again Figure 0. dA = C dr dt dt *Similar considerations underlie the basic integral theorems of vector analysis, a classical subject notoriously obscured by modem rigorous treatments. The outstanding text by Schey cited below presents a lucid informal account of those topics from a physicist's point of view, while vividly describing the bewilderment and hostility that such material often evokes from students previously exposed to abstract presentations. reasonable to conjecture that the rate should be proportional to the length of the curve, but exploration of the assumptions involved could quickly lead to such deeper concepts as rectifiability, orientability, and homology. (As in any discovery approach, the instructor must be well-grounded in his discipline to anticipate and lead discussion of such matters!) In any case, we are here concerned with the less ambitious aim of conjecturing the product rule. Toward that end, let h(t) and w(t) denote the height and width, varying as functions of time t, of a rectangle having one pair of adjacent sides lying along the axes of a Cartesian coordinate system. Any product may be represented as the (signed) area of such a rectangle, and we may think of the sides not lying along the axes as forming an L-shaped wavefront. By analogy with the ripple, at any time t the side of length w(t) will be moving with speed ht(t), and the side of length h(t) will be moving with speed w'(t) (Figure 1). Figure 1. [ h ( t ) w ( t ) r = h l ( t ) w ( t ) + h ( t ) w f ( t ) . Hence the total rate of engulfing (or disgorging) of area should be w(t)h'(t) h(t)w'(t). To prove the correctness of this result, we draw the rectangles corresponding to times t and t + At. In case all of h(t), hf(t), w(t), and wf(t) are positive, we obtain Figure 2, in which the L-shaped region representing the difference of the areas (numerator of the difference quotient) is left unshaded. + h(t + At) w(t) Figure 2. A ( t + A t ) - A(t) = h(t + At)w(t + A t ) - h ( t ) w ( t ) . There are two quite natural ways (Figures 3 and 4) to decompose that region into a pair of rectangles, each yielding an alternative form of the identity invoked in the algebraic proof. Of course, there are other cases to consider, and it would be well for students to draw the figures corresponding to other sign combinations; nonetheless the algebraic proof is perfectly general. The illustration (Figure 5) for the special case is also enlightening, since it is obvious that if one dimension is held fixed, the area of the rectangle can still change through the independent motion of the other dimension. w(t + At) h(t + At) - h ( t ) + At) - w ( t ) A ( t + At) - A ( t ) = [ h ( t + A t ) - h ( t ) ] w ( t + At) + h ( t ) [ w ( t + A t ) - w ( t ) ] . w(t Figure 3. h(t Figure 4. A ( t + At) + At) - A ( t ) = [ h ( t + At) - h ( t ) ] w ( t )+ h ( t + At)[w(t + A t ) - w ( t ) ] . w(t) Figure 5. [cw(t)l' = cwf(t). Two further observations are in order. First, in seeking to discover the form of the product rule, one may wish to consider examples of functions f # F and g # G for which f' = F' and g' = G' but (fg)'# (FG)'; such examples show that no formula for (fg)' involving only f' and g' can possibly be correct, thereby providing a simple demonstration of a non-existence theorem. Second, box diagrams are useful in many other contexts-in the author's opinion, their range of applicability is at least as wide as the much-touted Venn diagrams. As an example, consider Bayes' theorem. The formula P(A ( B )= P(A)P(BIA) P(A)P(B I A) + P(S - A)P(B I S - A) corresponding to events A, B in a sample space S is difficult for most students to remember, and its analytic derivation means little, if anything, to them. Yet after struggling mightily in vain attempts to train students to solve Bayesian problems, the author was amazed to discover how readily the ideas were assimilated when they were presented in the context of a box diagram rather than a formula. For example, consider the twins problem: Roughly 3/4 of all twins are fraternal, while only 1/4 are identical. Among identical twins, half are girls and half boys, while among fraternal twins, half are mixed sets and 1/4 each consist of just boys or just girls. Given that a set of twins consists of two boys, what is the chance those twins are fraternal? Many students never learned to fit such a problem into Bayes' formula, but after a single lecture most could quickly draw the diagram below and use it to obtain the solution. Many even remarked how easy such problems were to solve! 3/4 Fraternal 1/4 Identical 3/16 Figure 6 . P(Fraternal1 both boys) = 3/16 + REFERENCE H. M. Schey, Div, Grad, Curl, and All That, Norton, 1973. 3 = 5
© Copyright 2026 Paperzz