Of Camels, Inheritance, and Unit Fractions
Transcription
Of Camels, Inheritance, and Unit Fractions
Of Camels, Inheritance, and Unit Fractions A Paul K. Stockmeyer lawyer friend, who is also a mathematics enthusiast, recently related to me this story that had been a part of a sermon at his church. Under the laws of the country in question, inheritance was as follows. The eldest son took 1/2, the second son took 1/3, and the third son took 1/9. A man died, survived by three sons, and owning 17 camels. The sons began to bicker, because the law, if applied to the camels, would not be good for the camels, or the sons. Finally, they went to the village priest for advice. The priest listened to the story, and said, “Look, I have a camel. Take it. Then you’ll be able to work things out.” So the brothers took the priest’s camel, and divided up the camels. The eldest son took 1/2 of the 18 camels, or 9 camels. The second son took 1/3, or 6. The third son took 1/9, or 2. And much to their surprise, they had a camel left over. So they took it and gave it to the priest. Readers may be familiar with this story. According to recreational mathematics historian David 8 September 2013 : : Math Horizons : : www.maa.org/mathhorizons mh-08-11-stockmeyer-r2a.indd 2 8/16/13 4:56 PM Singmaster in his unpublished work Sources in Recreational Mathematics, problems of proportional division date at least as far back as the Rhind Papyrus (around 1650 BC), where an example is given of how to divide 700 loaves in proportion 2/3 : 1/2 : 1/3 : 1/4. However, the mathematical recreation involving the loan or donation of an extra beast to avoid the subdivision of individual animals dates only from the late 19th century. The first appearance of this type of puzzle that Singmaster records is in the book Hanky Panky: A Book of Conjuring Tricks, edited by William Henry Cremer and published in London in 1872. No author is listed, but the book is often attributed to the German magician Wiljalba Frikell. Others claim it is by Henry Llewellyn Williams. The book describes “a Chinese puzzle” of dividing 17 elephants into parts 1/2, 1/3, and 1/9. The puzzle appeared in several other books and magazine articles before the end of the 19th century, including those of the great American puzzle master Sam Loyd and the British puzzle king Henry Dudeney. Most writers reproduced the 1/2, 1/3, 1/9 version, with 17 animals of some sort, but a few other versions also appeared. Most include an imaginative setting in ancient Arabia or Asia. Preliminary Explorations As soon as the church service was over, my lawyer friend met with the church organist, another mathematics enthusiast, to discuss variations. Can the number of camels be, say, 19? or 34? or 35? Are at least three heirs required? What about four? Do we need to change the fractions? They soon contacted me for help. Can something like this be done with 19 camels? Yes. If the inheritance shares are 1/2, 1/4, and 1/5 (summing to 19/20) and the estate is 19 camels, then the donation of one more camel yields an augmented estate of 20 camels, from which the sons receive 10, 5, and 4 camels, respectively, leaving one camel to return to the donor. Is there a similar scenario with 34 camels? In the next section, we illustrate a method for showing that the answer is no. For an estate of 35 camels, readers can confirm that estate shares of 1/2, 1/3, 1/12, and 1/18 will divide up properly after borrowing a 36th camel. Four other schemes are also possible with 35 camels. But what exactly is it that we seek? An important part of mathematics is constructing appropriate definitions, the rules by which we play the game. We start this process by observing properties that all these scenarios have in common: • There is always exactly one borrowed camel added to the estate to form an augmented estate. The distribution of the augmented estate is completed with the return of the borrowed camel to the lender. • The estate shares sum to less than 1. • Estate share fractions are applied to the augmented estate, not to the original estate. • The estate shares are all distinct, with older sons inheriting more than younger sons. • The estate shares are fractions with 1 as the numerator. (Such fractions are called unit fractions, or Egyptian fractions. The ancient Egyptians are reported to have written all fractions between zero and one as a sum of distinct unit fractions.) • No camels are butchered in the distribution of estate shares. Each share of the augmented estate is an integral number of camels. Getting Formal Suppose we want to find an inheritance situation with sons and an estate of camels. To do so, we need to find a representation of 1 as a sum of unit fractions, fractions for the estate shares of the sons and one for the donor. The augmented estate will consist of camels, and the donor’s share is to be one camel, so the final fraction must be To avoid cutting up camels, all shares must be multiples of the one-camel share of In other words, all shares must be able to be written as unit fractions where the denominator is a divisor of Finally, we want essentially distinct unit fractions. We will allow one exception: the last son’s share can be one camel, so the last two fractions can both be All other fractions should be distinct. The decisions made in the preceding paragraph are somewhat arbitrary, and someone else might frame things differently. Nonetheless, these definitions and concepts are important, and we repeat them with proper formality. Definition: A distinct-share inheritance situation with sons and camels consists of a representation of 1 as the sum of unit fractions, listed in decreasing order of fraction size, with the following properties. Illustration by Gregory Nemec www.maa.org/mathhorizons : : Math Horizons : : September 2013 9 mh-08-11-stockmeyer-r2a.indd 3 8/16/13 4:56 PM 1 2 1 3 1 9 1 18 1 1. The last (smallest) fraction must be 2. The denominators of all the fractions must be divisors of 3. The fractions must be distinct, except that the last two fractions can both be The simplest situation is This corresponds to an inheritance with son and an estate of camel. The augmented estate is two camels, and the son and the donor each receives half, or one camel. There are two cases with two sons. The first representation is with an estate of five camels. The augmented estate is six camels, so the first son gets three, the second son gets two, and the donor gets back one. The other case with two sons is where an estate of three camels is augmented to four. Here the first son gets two camels, while the second son and the donor each gets one. The table at right lists the seven situations for three sons. (Remember, the fourth fraction is for the donor.) The third row shows the original situation. There are 52 possible situations with four sons. They range from with 1,805 camels down to with 15 camels. The preparation of full and accurate listings for more than four sons would certainly be aided by the use of a computer. Those with programming skills might 1 2 1 3 1 12 1 18 1 36 Estate Shares 1 Camels in Estate 41 23 17 11 19 11 7 enjoy designing an algorithm to produce such listings. Rather than organizing situations by the number of sons, we could try organizing by the number of camels in the estate. We have seen situations with 1, 3, 5, 7, 11, 15, 17, 19, 23, . . . camels so far. What about 9 camels? 13 camels? 21 camels? Are there any situations with these numbers? The answer is no. Suppose we had an estate of 9 camels. Then the augmented estate would contain 10 camels, and the last term in the summation would be 1/10. The denominators of the other terms of our 10 September 2013 : : Math Horizons : : www.maa.org/mathhorizons mh-08-11-stockmeyer-r2a.indd 4 8/16/13 4:56 PM summation must all divide 10. The most we can get is which is short of 1. With 13 camels, the only allowable terms in the summation are which again is short of 1. The case of 21 camels is similar. With C camels, we need to have lots of divisors so that there are lots of fractions to choose from in our summation to 1. With 9, 13, or 21 camels, we see that the numbers 10, 14, and 22 each has only three usable divisors, and that’s not enough to get fractions adding to 1. The numbers 10, 14, and 22 are all examples of so-called deficient numbers—a positive integer n is deficient if the sum of its divisors (excluding n itself) then is less than n. If the divisor sum is exactly n is said to be almost perfect. If the divisor sum is exactly n, then n is said to be perfect. Knowledge of perfect, almost perfect, and deficient numbers is helpful in this investigation. Questions to Explore Our results suggest many avenues for further discovery. Here are a few relatively easy questions: 1. Are there any interesting infinite families of distinct-share inheritance situations? In particular, is there such a situation for every positive number of sons? 2. Are there any distinct-share inheritance situations in which the eldest son’s share is 1/3, rather than 1/2? What about 1/4, or 1/5? 3. Are there any situations where the estate consists of an even number of camels? 4. Is there always a situation with for any Anyone who plays with these questions for very long will quite likely discover additional interesting questions to explore. There are other variations that I find interesting. The first appears in the book Puzzles for Puzzlers by Jonathan Always, published by Tandem in London in 1971. In this version, the estate is 13 camels, and the sons are to receive shares of 1/2, 1/3, and 1/4. Note that here the shares add up to more than 1. The lawyer handling the estate first takes one camel as his fee, leaving a diminished estate of 12 camels to be distributed. He then dictates that the first son should receive camels, the second son should receive camels, and the third son should receive camels. But this sums to 13 camels, and the diminished estate has only 12. The lawyer is forced to return the 13th camel, and everyone else lives happily ever after. Finally, there is a variation in the book O Homem One of my mentors claimed that research is what we engage in when we don’t know the solution to a problem. Inheritance questions are wonderful areas where almost anyone can enjoy doing research. que Calculava by the Brazilian writer Júlio César de Mello e Souza, writing under the pen name Malba Tahan. An English translation called The Man Who Counted was published by Norton, New York, in 1993. This book is a sort of recreational mathematics version of the Arabian Nights. In this variation there are 35 camels, with the usual shares of 1/2, 1/3, and 1/9, but with both a lawyer and a priest! The lawyer gets the priest to donate a 36th camel, and then declares that the sons should receive camels, camels, and camels. This sums to 34 camels, leaving one to return to the priest and one for the lawyer to keep as his fee. Readers are encouraged to explore one or both of these variations. Start by gathering data—find other scenarios similar to one of these. List what these scenarios have in common, then create a formal definition of an inheritance structure of this type. Continue to explore this new structure as we have done here with the original. Are there infinite families of numbers for which you can construct valid situations? Can you rule out infinite families of numbers? One of my mentors claimed that research is what we engage in when we don’t know the solution to a problem. Inheritance questions of the type discussed here are wonderful areas where almost anyone can enjoy doing research! n Paul Stockmeyer is a professor emeritus of computer science at the College of William and Mary. His retirement activities include serving as an associate editor of Mathematics Magazine, teaching mathematics-related courses for senior citizens, and puttering around in recreational mathematics. Email: [email protected] http://dx.doi.org/10.4169/mathhorizons.21.1.8 www.maa.org/mathhorizons : : Math Horizons : : September 2013 11 mh-08-11-stockmeyer-r2a.indd 5 8/16/13 4:56 PM