Every earlier rate in this Subject came out as a fraction, because the functions were polynomials and the quotient was a polynomial in h. sin and cos are not polynomials, and at almost every place the value is irrational, so no column of fractions is available. The lab here computes the quotients in ordinary floating point, prints each rounded to six significant figures behind ≈, and prints the number they head for. The honest status of that last number is the important point, and it is stated below. The lab's fourth tile divides the last quotient by f(a); it was built for the exponential, whose rate is a fixed multiple of its value, and for sine and cosine it says nothing, so the first three tiles are the ones to read.
Theorem · The rates of sine and cosineFor t measured in radians, the rate of sin at every t is cos t, and the rate of cos at every t is −sin t. In symbols: sin′ t = cos t and cos′ t = −sin t.
This is a claim, and the lab demonstrates it at the places you choose without proving it. A table of quotients that settle toward a number shows the pattern; it cannot show that the pattern continues for every smaller h, and no table can. The course uses the claim from here on, as every course in this Subject uses the rate of an exponential, and says so once.
The column at zero
At a = 0 the first sine preset starts with h = 1. The quotient is (sin 1 − sin 0)/1 = sin 1, which is ≈ 0.841471 and has no fraction form. Halving h gives ≈ 0.958851, ≈ 0.989616 and ≈ 0.997398, and the lab continues to the sixth halving. The numbers climb toward 1, which is cos 0, and the gap to 1 is cut to about a quarter each time h is halved. A quarter, and not the half that the cosine table below will show, because the leading error of a one-sided quotient is proportional to h times the second rate of the function at a, and sin″ 0 = −sin 0 = 0: at 0 that term is absent and the next one, proportional to h², is what remains.
h = 1 ≈ 0.841471
h = 1/2 ≈ 0.958851
h = 1/4 ≈ 0.989616
h = 1/8 ≈ 0.997398
claim: these head for cos 0 = 1
The sign of the cosine's rate
The cosine preset at a = 0 is the one where the sign hides. Its quotients start at ≈ −0.459698 and shrink toward 0, because cos is at the top of its hill at 0 and is not yet falling. That does not tell you the sign of cos′ in general, since 0 has no sign. The preset at a = 1 does: its quotients are all negative and head for −sin 1 ≈ −0.841471, because the cosine is falling at 1. A reader who remembers only that the rate of cos is “the other one” will write sin 1, and the table refutes it with a minus sign.
Example · Reading the cosine at 1The first quotient, h = 1, is cos 2 − cos 1, which the lab prints rounded as ≈ −0.956449. It is negative because cos is smaller at 2 than at 1.
The last of the seven quotients, at h = 1/64, is ≈ −0.845658. It is still moving toward the limit −sin 1 ≈ −0.841471, and from the third row on the gap shrinks by about a half each time, which is what a quotient with an error proportional to h does.
Radians matter. A degree is π/180 of a radian, so a table built on degrees would head for cos times π/180 instead, and the rates would not be each other's. Everything here, and the whole course, measures the argument in radians.
Theorem · Second rates of sine and cosinecos″ t = −cos t and sin″ t = −sin t for every t.
Use the two claims above. The rate of cos is −sin, so the rate of the rate of cos is the rate of −sin, which is −(sin′), and sin′ = cos. Hence cos″ = −cos.
The same steps give sin″ = (cos)′ = −sin. The proof is only as firm as the two claims it uses, which the table demonstrates and does not prove.
□
Both identities say that the second rate of the function is minus the function. Written as an equation for an unknown y, that is y″ = −y, or y″ + y = 0. The next lesson substitutes into it and finds that cos, sin and every combination of them solve it.