Al-Samaw'al's curious approach to trigonometry (Q2871012)

For Ptolemy and the astronomers who came later, a table of sines containing the values of \(\sin n^{\circ}\) for every \(n \in \mathbb{N}\) is understandably indispensable. Such tables were usually constructed using the method that appears in Ptolemy's \textit{Almagest}, and that usesNEWLINENEWLINE(i) the addition formulas for \(\sin (A\pm B)\),NEWLINENEWLINE(ii) the half-angle formula for \(\sin (A/2)\), and the interesting inequalityNEWLINENEWLINE(iii) \(\sin A^{\circ} / A < \sin B^{\circ} / B\) for \(0 < B < A < 90\).NEWLINENEWLINENEWLINENEWLINEIt is obvious that once \(\sin 1^{\circ}\) is calculated, \(\sin n^{\circ}\) can be calculated for all \(n \in \mathbb{N}\) using (i) repeatedly. Also if an angle \(A^{\circ}\) is constructible (by straightedge and compass), then its sine can be expressed in terms of square roots, and hence can be satisfactorily approximated. Since the regular pentagon and hexagon are constructible (as shown in Book IV of Euclid's \textit{Elements}), it follows that one can calculate the sines of their central angles \(72^{\circ}\) and \(60^{\circ}\), hence the sine of \(72^{\circ} - 60^{\circ} = 12^{\circ}\) using (i), and then the sines of \(6^{\circ}\), \(3^{\circ}\), \(1.5^{\circ}\), \(.75^{\circ}\), and so on, using (ii). One can never reach \(1^{\circ}\) in this manner (or by any other method, since we know now that \(1^{\circ}\) is not constructible). For \(\sin 1^{\circ}\), Ptolemy and later astronomers used the the inequality in (iii) which yields \((2/3) \sin (1.5^{\circ}) < \sin 1^{\circ} < (4/3) \sin (.75^{\circ})\).NEWLINENEWLINEThe paper under review is concerned with an unusual table of sines that appears in Chapter 4 of a book entitled \textit{Exposure of the errors of the astronomers} written by the twelfth century Islamic scholar al-Samaw'al. In his table of sines, al-Samaw'al decided to break away from the tradition of dividing a straight angle into 180 degrees (i.e., \(180^{\circ}\)), and he divided it instead into 240 \textit{new} degrees (say, \(240^{ \diamond})\). Thus the central angles of the regular pentagon and hexagon would be \(96^{ \diamond}\) and \(80^{ \diamond}\); and \(1^{ \diamond}\) would be obtained by successive halving of the difference \(96^{ \diamond} - 80^{ \diamond} = 16^{ \diamond}\). Essentially, the unit is taken as \(1^{ \diamond} = .75^{\circ}\) instead of \(1^{\circ}\). A table of sines using the new unit \(\diamond\) would never need to use the inequality (iii)!NEWLINENEWLINEThe authors of the paper under review reproduce al-Samaw'al's table of sines, as well as the Arabic text and an English translation of Chapter 4 of the afore-mentioned book of al-Samaw'al. They point to errors in the table, and they make speculations on whether al-Samaw'al used other existing tables in constructing his own. They quote objections, made by the eleventh century mathematician and astronomer Ibn al-Haytham, to Ptolemy's use of (iii), and summarize the significant improvements on Ptolemy's tables made by the Islamic scholars Ibn Yūnus, Abū'l-Wafā', al-Bīrūnī, and al-Kāshī.NEWLINENEWLINEIt is worth mentioning that al-Samaw'al is best known for his \textit{Dazzling book of algebra} in which Pascal's triangle appears several centuries before Pascal and Newton. Al-Samaw'al's \textit{Pascal's triangle} is reproduced on the front cover of [Notices Am. Math. Soc. 60, No. 11 (2013)].