A proof of Goldbach's hypothesis that all even numbers greater than four are the sum of two primes. By Kent G Slinker Abstract In this paper I introduce a model which allows one to prove Goldbachs hypothesis. The model is produced by studying Goldbach partitions as displayed by an inverted mirror image of all the primes up to some even number equal to the last prime plus three. The bottom half of the model is then moved to the right in steps of two which exhibit the Goldbach partitions for the next even number. As long as the model contains all the primes up to the resulting even number minus three, then Goldbach’s hypothesis can be proven if it can be shown that each move must produce a Goldbach partition until one reaches the next prime plus one. I show that this must be the case. Notation k m p pa pn pn +1 pn −a ra rt Any even number A specific even number, always three more than some prime. A prime number (2, the only even prime, is not considered in this paper). Any prime in the list which makes up a Goldbach block. The last prime (also the greatest) in a Goldbach block. The next prime after pn . The mirror image counter-part of pa in a Goldbach block. The gap between pa and pn −a in a Goldbach block. An even number added to m to obtain k. Introduction Christian Goldbach (1690-1764) proposed, in a now famous letter to Euler in 1742, that every even number greater than four can be expressed as the sum of two primes. This was also part of Hilbert’s 8th conjecture at the 2nd International Congress in Paris in 1900. Hardy and Littlewood stated in a paper which appeared in Acta Mathematica in 1922, “There is no reasonable doubt that the theory is correct, and that the number of representations is large when m is large; but all attempts to obtain a proof have been completely unsuccessful.” Several computational results have shown that Goldbach's hypothesis is true for a large number of even numbers. Javier Echevarria verified the hypothesis up to 232 , and Matti Sinisalo to 4 x1011 (RG). Sheldon calculated that the probability of 1 some even number m > 4 x1014 not having a Goldbach partition to be approximately 10−150,000,000,000 (NS). Impressive as this may be, it is not a proof, and the possibility remains that there does exist some even number which is not the sum of two primes. The seeming random distribution of primes have made the conjecture quite difficult to prove, along with the lack of techniques related to the additive properties of primes. It is certainly possible that, even though very rare, there are an infinite number of exceptions to Goldbach's conjecture. The question is open as to why no proof of Goldbach's hypothesis has appeared. I suggest that the fault does not rest in the existing methods of analysis, but rather in the lack of a model to guide that analysis. If I am correct, and the following model presented in this paper does provide a proof of Goldbach's conjecture, then this will be no small victory for model theory and should reemphasize the need for the construction of models as an aid to existing techniques of analysis. The Model To prove Goldbach's conjecture I will employ a model based on the following simple equation: Where p is some odd prime, k an even number, and 3 ≤ p ≤ (k - 3): (k − p ) + p = k This equation supplies one of the necessary primes; if it can be demonstrated that (k − p ) can always be equal to another prime for every even number k, then the hypothesis is proven. Example: Take for instance the number 26. It has three Goldbach partitions: 23 + 3 = 26 19 + 7 = 26 13 + 13 = 26 For each prime in the partition: {3, 7, 13, 19}, the differences: {(26 - 3), (26 - 7), (26 - 13), (26 - 19} are also prime. We can now construct a model that expresses these partitions. We first write down each prime as a block up to (26 - 3), since 23 is prime it forms the top prime row. The top row contains the integers up to 26 for reference. 2 Figure 1 If we now mirror and invert the prime rows of this model, we can see the Goldbach partitions as listed above. Figure 2 Let us now move the right side of the mirror image 2 moves to the right. This will yield the Goldbach partition for (26 + 2): Figure 3 The resulting configuration will show all the Goldbach partitions for 28, which has only two: 5 + 23 , and 11 + 17 . Figure 3 shows these partitions along with their commutative equivalents: 17 + 11, and 23 + 5 . 3 We could move the Goldbach block two more moves to the right to obtain all the Goldbach partitions for 30, however, as soon as we move the block 26 + 6 we lose one Goldbach partition for 32, namely 29 + 3 . This is because our model does not contain all the primes up to 32 − 3 . If we restrict our moves to the right such that from the initial position at m every move to the right contains all the primes up to the resulting even number minus three, we can prove Goldbach’s hypothesis by showing that every such move must produce a Goldbach partition. Goldbach Blocks Since this entire analysis rests on what I have termed a “Goldbach block”, let us now define them. Each Goldbach block will start at some initial state represented by some even number m, and consist of two mirror images of the list of all primes {p0 , p1, p2 … pn } , with the first prime being p0 = 3 and the last prime, pn always equal to m − 3 (see Figures 1 and 2), hence: [Def. 1.0] The first prime p0 . p0 = 3 [Def. 1.1] The last prime in the Goldbach block pn . pn = m − p0 [Def. 1.2] The value of m at the Goldbach block’s initial position. m = p0 + pn For example, in Figure 4 below, the prime list goes from { p0 , p1, p2 … p7 } , so that pn = p7 which has a value of 23. The value of p3 is 11, and the value of p7 −3 = p4 is 13. The value of m is p0 + pn , or 3 + 23 = 26 . We will now designate the "gap" between each prime, pa , and its mirror image counter part, pn-a , as ra at the Goldbach block's initial configuration at m. Figure 4 shows the different ra as applied to Figure 2: 4 Figure 4 From the fact that m = pn + p0 , r0 is always equal to zero. The value of each individual ra can be seen to be: [Eq. 1.0] The value of each gap in the Goldbach block. ra = m − ( pn −a + pa ) , or solving for m: [Eq. 1.1] m = ra + pn −a + pa Example: In the above figure m = 26 , p2 = 7 , and p7-2 = 17 , hence r2 = 26 − (17 + 7 ) = 2 . By definition, Goldbach blocks at the initial configuration at m have at least one Goldbach partition, since m = p0 + pn . Let us examine a Goldbach block with the right hand mirror image shifted to the right 4 spaces to obtain m + 4 . Figure 5 shows this configuration: Figure 5 5 Since r0 = 0 by definition, we designate rt as the number of moves to the right from the initial position at m = pn + p0 , hence rt at m is zero. We will define the new sum, represented by m + rt as k, hence: [Eq. 2.0] The value of a Goldbach block shifted to the right. k = m + rt Where m + rt ≤ pn +1 + 1 Solving for rt in terms of k and m: [Eq. 2.1] k − m = rt Where rt ≤ ( pn +1 + 1) − m Definition of a Goldbach Partition Let us now define a Goldbach partition for these Goldbach blocks. Since a partition occurs when a prime in the bottom mirror image matches a prime in the top image, then if some prime ( pa ) in the prime list plus the original gap ( ra ) plus the move to the right ( rt ) equals another prime higher up on the list, then the resulting Goldbach block exhibits a Goldbach partition. Let pm be another prime in the prime list { p0 , p1, p2 … p7 } such that pm > pa then: [Theorem. 1.0] Definition of a Goldbach Partition If pa + ra + rt = pm , then k has a Goldbach partition equal to pm + (k − pm ) . An easy Corollary to [Theorem 1.0] is: [Corollary. 1.1] If k − ( pa + ra + rt ) = pq , then k has a Goldbach partition equal to pq + (k − pq ) Proof: From [Theorem 1.0], pa + ra + rt = pm which allows us to re-write [Corollary 1.1] as k − pm = pq , which leads immediately to k = pq + pm Example: Specifically, in the above example for 30 (Figure 5), one Goldbach partition is 11 + 19 . Hence from Theorem 1.0: 6 pa + ra + rt = pm Instantiating pa + ra → p1 + r1 p1 + r1 + rt = pm Replacing each variable with the values given in Figure 5: 5 + 2 + 4 = 11 Hence, from [Theorem 1.0], the Goldbach partition for 30 is given by: 11 + (30 − 11) = 30 , or: 11 + 19 = 30 Proof of Goldbach’s Hypothesis We can now prove that each k must have a Goldbach partition. Proof: From [Eq. 2.0] k = m + rt If rt = 0 , then k = m , and hence has a Goldbach partition from [Def. 1.2] Suppose rt ≠ 0 , then from [Eq. 2.1] k − m = rt But from [Eq. 1.1] m = ra + pn −a + pa So we can write: k − ( ra + pn −a + pa ) = rt Rearranging terms: k − ( pa + ra + rt ) = pn −a 7 But from [Corollary 1.1], if k − ( pa + ra + rt ) = pq , then k has a Goldbach partition equal to pq + (k − pq ) , hence k has a Goldbach partition at pn −a + (k − pn −a ) . Kent Slinker Pima Community College kslinker@pima.edu 8 References [NS] NEIL SHELDON. A Statistician's Approach to Goldbach's Conjecture. Teaching Statistics. Vol. 25, No. 1. Spring 2003. [RG] RICHARD K. GUY. Unsolved Problems in Number Theory. Springer-Verlag. New York, 1994. [WY] WANG YUAN (editor). Goldbach Conjecture. World Scientific Publishing Co Pte Ltd. Singapore, 1984. 9