āChapter 1
Of Sand and Stars
Are there more grains of sand on Earth or stars in the universe? With your eyes alone you can see at least a couple of thousand stars on a clear night well away from artificial lights, and nearer 4,000 if itās moonless and your eyesight is keen. In a handful of sand are many more grains than that. But space is huge, dauntingly so, and powerful telescopes reveal that it contains a host of galaxies, each harbouring billions of stars. On the other hand, the deserts, beaches and ocean beds of our planet are home to sand particles in dizzying profusion. So, sand or stars, which wins in the numbers game?
A study carried out by researchers at the University of Hawaii in 2003 estimated the number of sand grains in the world to be 7.5 million trillion, or 75 followed by 17 zeros. As for stars, the figure they came up with, for the whole of the observable universe, was 70 thousand million trillion. Thatās about ten thousand stars for every sand grain.
The Greek mathematician and scientist Archimedes was also interested in this kind of problem. In the third century bce he wrote a short treatise, addressed to Gelon, King of Syracuse, thatās come to be known as āThe Sand Reckonerā. Sometimes described as the first research-expository paper, because it combines both accuracy and clear language, aimed at the layperson, it asks: How many grains of sand would fit in the universe?
Figure 1.1: Sand dunes in the Sahara, Libya.
The answer, of course, depends on how big is an average grain of sand and how big is the universe. Archimedes figured, very generously (to the point of being unrealistic), that one poppy seed could contain 10,000 grains of sand, which would make the grains almost microscopic in size. He also reckoned that 40 poppy seeds, side by side, would stretch across one Greek dactyl, or finger-width, equal to about three quarters of an inch (19 millimetres). A sphere one dactyl wide would then be able to hold in the region of 640 million sand grains.
As for the size of the universe, Archimedes based his estimate on the classical heliocentric theory of his predecessor Aristarchus. In this model of space, Earth orbits around the Sun while the stars are fixed to a sphere, also centred on the Sun, but much further out. The fact that the Greeks couldnāt discern any change in the relative positions of stars in the sky ā a so-called parallax ā as Earth moved from one side of the Sun to the other meant that stars had to lie a certain minimum distance away. This gave Archimedes his estimate for the smallest possible diameter of the then-known universe ā in modern units, about two light-years.
Today we can easily do the maths and arrive at how many Archimedean-sized sand grains would fit inside a ball two light-years wide. The answer comes out to be roughly one followed by 63 zeros, which can be written compactly as 1063 ā meaning 10 Ć 10 Ć 10 Ć ā¦ Ć 10 (with 63 tens). The problem Archimedes faced is that our handy ways of representing big numbers didnāt exist in his day. The Arabic numerals, 0 to 9, that we now use, emerged about 800 years later (and in India, not Arabia). Place-value notation, in which the same symbol is used to represent different orders of magnitude depending on its position (for example, the ā3ā in 30, 300, and 3,000) was still in its infancy in Babylon but hadnāt yet reached Greece. And there was in those days no such thing as index notation, in which how many times a number must be multiplied by itself is written as a superscript (as in 1063).
At the time when Archimedes began his cosmic sand calculations the Greeks used letters of the alphabet to represent numerals. A different letter stood for the equivalent of our numbers 1 to 9, multiples of ten from 10 to 90, and multiples of a hundred from 100 to 900. The familiar 24 letters, alpha to omega, which have survived in present-day Greek, had to be supplemented by others taken from older languages and dialects to provide enough labels. Alpha to theta stood for 1 to 9, iota to koppa (borrowed from the Phoenician) for multiples of ten from 10 to 90, and rho to sampi (used in some eastern Ionic dialects) for multiples of a hundred from 100 to 900. The Greeks didnāt use the same letter again and again in different positions, so that, for example, 222 would be written as ĻĪŗĪ² (sigma kappa beta = 200 + 20 + 2). For multiples of a thousand, from 1,000 to 9,000, some of the same letters were employed but with various extra marks. And that was as far as the ancient Greek labelling system of numerals went, except for the murious ā the largest single unit defined, written as a capital mu (M) and equivalent to our 10,000. The Romans called it the myriad, a name that became absorbed into English but with the altered meaning of ācountlessā or a very large (but undefined) number.
The Greeks could write numbers that were bigger than a murious but only as multiples of M using strings of letters in the manner described. For example, 1,234,567 would be written as ĻĪŗĪ³Ī ĶµĪ“ĻĪ¾Ī¶ (123 Ć 10,000 + 4,567). Itās an approach that quickly runs out of steam for anything beyond what we would call a few hundred million.
Archimedes realised that to represent the kind of gigantic numbers that would arise from his cosmic sand calculations, heād have to come up with a whole new system of number naming. He started by defining anything up to a myriad myriad as being a number of the āfirst orderā. To us that mightnāt seem like a big step because we can easily write a myriad myriad as 104 Ć 104, which equals 108 (a hundred million), and then carry on indefinitely from there. But there was nothing like our index notation, in which an index or exponent is used to show how many times a number must be multiplied by itself, when Archimedes took on his big-number project.
Having defined any number up to a myriad myriad as belonging to the first order, he moved on to numbers that lay between a myriad myriad and a myriad myriad times a myriad myriad (1 followed by 16 zeros, or 1016 in modern notation). These, he said, belonged to the āsecond orderā. Then he progressed to the third order, and the fourth, and so on, in the same way ā each successive order being a myriad myriad times larger than the numbers of the previous order. Eventually, he reached numbers of the myriad myriadth order, in other words, in our index notation, 108 multiplied by itself 108 times, or 108 raised to the power 108, which equals 10800,000,000. All these numbers, of which the largest would have 800 million digits if written out in full, he defined as belonging to the āfirst periodā. The number 10800,000,000 itself he took to be the springboard for the second period, at which point he began the process all over again. He defined orders of the second period by the same method, each new order being a myriad myriad times greater than the last, until, at the end of the myriad myriadth period, heād reached the colossal value of a myriad myriad raised to the power of a myriad myriad times a myriad myriad, which weād write as 1080, 000, 000, 000, 000, 000, or 10 to the power of 80 thousand trillion.
Remember, Archimedes had no knowledge of our compact ways of writing big numbers. There wasnāt even the concept of zero in ancient Greek maths. Starting from a system that struggled to name numbers that were bigger than a few hundred million, he fashioned a method to describe a number that, in decimal form, would have 80 thousand trillion digits.
For the purposes of his sand-counting project, it turns out, Archimedes didnāt need numbers anywhere near this large. Using his estimates of the size of a grain of sand and of the whole universe, he came up with a value that was only of the eighth order of the first period. In index notation, a mere 8 Ć 1063 or so of Archimedesā minuscule grains would have been enough to pack the two-light-year-wide Greek cosmos full of sand. Even using a modern, and much larger, estimate for the diameter of the observable universe of 92 billion light-years thereād be no room for more than about 1095 sand grains ā still a number of just the twelfth order, first period.
āThe Sand Reckonerā was cutting-edge stuff. Not only did Archimedes offer a picture of the universe that most closely resembles what we know today, given the limited data he had available, but he invented a whole new way of describing big numbers. He was the first person to tackle the problem of naming and manipulating large numbers without the benefit of modern notation. Using a system with base 10,000, he effectively pioneered exponentiation ā the process of raising one quantity to the power of another. He also discovered the law of adding exponents, namely xm Ć xn = xm+n, for any numbers x, m, and n; for example, 32 Ć 33 = (3 Ć 3) Ć (3 Ć 3 Ć 3) = 35.
Archimedes was the first person to show that itās possible to go beyond the tradition of his era of simply calling huge numbers of things āinnumerableā. Sand and stars, in particular, came in a lot for this kind of treatment. The Greek poet Pindar, who predated Archimedes, wrote in his Olympus Ode II: āsand escapes countingā. Thereās even a Greek word, psammakĆ³sioi ā literally, āsand-hundredā ā thatās used to mean āuncountableā. Writers of the Bible, too, gave up on the arithmetic of sand and stars. The phrase in Genesis (32:12) āthe sand of sea, which cannot be counted for multitudeā is one of twenty-one Biblical references suggesting that itās impossible to put a figure on the numbers of sand grains out there. Hebrews (11:12) conflates the two: āSo many as the stars of the sky in multitude, and as the sand which by the seashore is innumerable.ā
As weāve seen, Archimedes didnāt confine himself to the sand on a seashore or even on the Earth as a whole. He made sure that none of his contemporaries could possibly outdo his number count by imagining the entire universe to be packed full of sand grains so small that theyād be barely visible. It would be interesting to know what heād have thought of the efforts of other intellectuals, a few hundred years later, who also wrote about large numbers but in a different part of the ...