Login Form

The Divine Gardener

Author: Eli Ehrman, Adar 5769/March 2009

Talk given by Eli Ehrman on the occasion of the Yahrzeit of his sister Etta Z"L

I want to propose a simple idea – not much more than a thought-bite. The idea itself can be explained quite briefly. Proving it takes a little longer. The implications and applications of the idea, however, can go a long way.

The idea is this: G-d is not a super-engineer, a master designer nor an architect on a universal and microscopic level. At least, that model of the way He works is neither accurate nor productive. Instead, a more correct image of G-d is that of a Cosmic Gardener.

A gardener creates within the restrictions of the laws of nature. G-d may differ from the human gardener in that He authored the laws and is bound by them only by choice. Nevertheless, he has formed the Universe since its first moment of inception largely as if He was indeed constrained by such laws.

However, the critical difference between an engineer and a gardener is that the perfect engineer should experience no surprise. The Perfect Designer would plan out His Creation down to the minutest detail and then implement it exactly according to the plan. The phrase "and he saw it was good" would not be applicable for such an ideal architect.

A gardener may guide, uproot, nurture and weed his creation but the initiative seems to come from the plants themselves. The gardener may have an overall criterion for what is good: He may desire that the garden be green and lush or colorful and diverse. However, the exact dimensions of the outcome are not determined in advance. Rather, if a beautiful plant does well and flourishes, such a plant will be given attention and help. It may be given extra space or encroaching intruders weeded. A gardener may plant a specific species but since much depends on the seemingly independent growth forces of the plants themselves, the gardener may very well be surprised, reflect on the outcome and see that it is good. Each individual plant may develop in ways not under the direct control of the gardener. The plant as such proposes. The gardener, in response, disposes. A gardener is, in essence, an opportunist.

I propose a model that unifies G-d's guidance of His Creation both during the pre-human evolutionary phase as well as the human, historical period. We are used to the idea that human beings may have Free Choice. However, it is not only human beings that have the power to propose and create new historical directions. Life as a whole is allowed to develop its own paths. Nevertheless, in both phases there is a complex and dynamic relationship of initiative from one side and guidance from the other. There is outburst form one side and prevention or limitation from the other. In neither case is the outcome determined in advance. There are always many possible outcomes. Nevertheless, there are clear criteria for what the nature of the results should be. There is always a distinction between desirable and unwanted.

Combining and unifying two very different periods: the evolutionary and the human, gives us great intellectual power. We can apply the understandings and empirical observations of one to propose understandings of the other. We start therefore by showing that the gardener is the most appropriate model to apply to the process of evolution.

We need to start with some probability theory. We will try and keep the discussion intuitive and concrete. We can get the ideas across without abstract mathematical formulations. Let us start with a simple thought experiment. Take three coins. If we flip them all, there is only a one in eight chance that they will all end up heads up. That means that we need to repeat the experiment about eight times to get a reasonable chance that we should succeed at least one time. If we have 4 coins we need 16 times, 5 needs 32 and so on. If we have one thousand coins we would need an astronomical number of times: we would write it as a 1 with about 300 zeros after it. The thirteen billion years that our Universe has existed for is a number of seconds that is less than 1 with 18 zeroes after it. So even if we repeated the experiment once every second since the Universe began we would not have a hope. There are, as far as Cosmologists know, at most trillions of stars in the Universe. Even if they each have lots of planets and we repeat the experiment on every square meter of every planet every second we would not get close.

However, all hope is not lost. There is a way to get all 1000 coins to end heads up and it's going to take less than ten tries! I flip all 1000 coins. Around 500 will be heads up. I hold these down in the next try and allow only the rest to flip. Now another 250 will be heads up and so in total around 750 are heads up. I hold all these 750 still and allow only the rest to make their random moves. We will now have 875, then 937, 961,985,993,996,998,999 and then all 1000 will be heads up. The last few numbers are a little uncertain because the statistical sample is low but by about 15 tries there is very little chance of not having achieved our goal completely. We have made the impossible easy!

Now you might say that I might as well simply set up all the coins heads up and forgo the whole random pretence. However, the fascinating thing is that, as I will show soon, G-d or whoever the Cosmic Designer is, seems to use exactly this method. He does not just build some fixed outcome ex-nihilo. Instead, what we see is exactly such a process of random fluctuations that nevertheless make progress due to continuing intervention and guidance.

Why would the Cosmic Designer or Cosmic Gardener work in this way? While empirical observations should not immediately be translated into explanations, there is more sense to such a method than is implied by the coin example. The coin example presents only two stark choices for each coin and the end result is very fixed and invariant. However, if there were more than one option for each element and if we did not require all the elements to go to only one solution then we would see the openness in the system.

We can illustrate this with a slightly modified thought experiment. Assume we have 320,000 or so pixels instead of 1000 coins. Each one can flip to any of three colors. We arrange these as you would on a VGA monitor with 480 rows of 640 pixels. If each second then flips randomly it is guaranteed that all you will have is a gray screen filled with what engineers refer to as "noise". However, if I let the pixels flip randomly but start applying rules such as holding still pixels that have neighbors of the same color I will suddenly have interesting shapes of different colors emerging on the screen. I can extend, manage and control such interventions to produce very pleasing aesthetic results. I would in this sense be acting as a gardener instead of a designer. Here is an example of random processes producing interesting outcomes that are surprising to the "gardener". There is an interesting interaction here between a form of initiative on the part of the random processes and the desires and control of the gardener.

The next example is more concrete. I take this example from what we understand of the actual process of evolution in human cognitive terms. I would like to compare the human brain with that of our nearest neighbor along the genetic evolutionary line: the chimpanzee. There are many, extremely complex innate advances present in the human brain relative to that of the chimpanzee. By innate, I mean a capability that it is not learnt but rather is found in essentially identical form in all human beings.

Now the human line diverged from that of the chimpanzees around 3-5 million years ago. Taking a generation to be about 25 years this represents no more than 200,000 generations. This is an absurdly small number of generations given the vast advances involved. However, that is just a psychological statement. A billion generations is also absurdly small but we tend to have more respect for numbers like a billion.

I use the example of the human brain because the advances we are talking about can easily be compared to computer programs or software even though the capabilities of the most complex human programs written to date don't even come close to those of the human brain itself. The simplest of our cognitive processes might be simulated by many millions of independent elements that need to be in exact order. This dwarfs even our 1000-coin example in terms of the impossibility of random process.

However, let's assume we can talk about building up a cognitive function out of very simple cognitive sub-functions that each confers the slightest tiny survival advantage to any ape that happens to have been blessed with such a capability. Let's further make the absurd assumption that such a minute cognitive capability – the smallest and simplest possible that still makes some difference can be represented by 1000 neurons wired in the right way.1

We already know that there is no chance on any planet in the universe that such a cognitive function should appear randomly. However, what if some ape is born with a few neurons in the right state. Now this neuron arrangement, or in fact, any of the incomplete sequence, is of no survival advantage. That is because we defined the complete sequence as the minimum to provide any survival advantage – even if it is as slight an advantage as possible. So there is no reason that this individual will survive rather than any other and so after a few descendents the configuration disappears from the gene pool. This is the equivalent of some coins flipping heads up but over time the coins are flipped again and rearrange.

However, instead, the Cosmic Gardener knows that the way to win the 1000-coin game is to prevent the coins that are already in the right configuration from staying in the random game of arrangement and rearrangement. These coins must be held still as the other coins keep flipping. So without any external survival advantage, the Gardener will see to it that this ape and his descendents will perpetuate these special genes. Over time other apes, whether descendents or potential mates, will randomly develop other pieces of the sequence. These too will be preserved. Eventually, the sequence will complete and one ape will gain the cognitive function. Of course, as we all know, some slight advantage is often blown away by the other, far more significant survival factors. Being a tiny bit more intelligent does nothing for you when you catch some variant of a terrible plague and die along with most of your tribe. However, in this case this new ape just will survive and prosper and another piece will fit into the overall cognitive constellation that we will eventually know as humanity.

This is the way a gardener would work. Something randomly shows up or does well and the gardener responds by nurturing a welcome innovation. This is opportunism rather than design. If something good comes out, the gardener goes with the flow and builds around it, making room for it and giving it a special chance.

But this is not a determinate process. Something starts developing and so it may be helped. On the other hand, the forces of random chance may just have their day instead. There are many ways to move forward cognitively. It might be one feature that turns up and it might be another. The goal is set that an intelligent creature that is aware of itself, the Universe and ultimately its Creator should emerge. Exactly how this should happen and what cognitive constellation this creature will have depends on what opportunities the random process throw up.

I don't know if I would have chosen this way of creating the Universe as we know it but, then again, nobody asked me. All we can say is that this seems to be the only way of explaining the evolution patterns that we observe. The outcomes are absolutely beyond random and yet it seems to build up bit by bit following opportunism generated by random processes. The conclusion that this is the way the Universe is guided is an observational, empirical one.

We learn from the observations of the early history of life as well as observations of the dynamics of genetics as it appears today that this is the nature of the behavior of the Creator.

The fascinating thing is that He seems to apply the same manner of guidance to human history too. This simple idea provides insight that is helpful to understanding why questions such as human suffering and why, in general human history seems to be the way it is. This is not the place to provide the detail but here too we propose that the right way to look at the problem is by starting out with the model of a Divine Gardener rather than a Divine Engineer. Slowly, ever so slowly, humanity emerges that learns to control more and more illnesses. Slowly we create a society where we can focus resource away from raw, brute survival. Who knows, perhaps we may even learn to translate this economic prowess into an existence focused on learning, advancement and self-understanding.

But this does not only provide a unified insight into pre-human and human history. This idea of G-d as a gardener provides insight into understanding what the Torah is telling us about His Guidance. This is also the image of G-d presented in the text of the Torah.

Throughout the Torah we see that G-d wants a specific goal: the good for all of His Creation. However, there are many paths that lead to that goal and presumably many forms that that good may take. We see Him respond again and again to the initiative and behavior of human beings. If they fail Him, then history takes one course and if they strive to create good in the world then history takes another. Human beings plead and pray and G-d responds. He even creates covenants with His creations agreeing to restrict Himself in a specific manner in return for them promising to keep the covenant.

We see that throughout the Torah, the actions of G-d revealed in the world follow the model of the responsive gardener far better than the model of the pre-ordaining engineer. G-d has run the world in the billions of years of its development in the model as He runs the affairs of humanity today.

Of course, the nature of the initiative is different. The initiative of human beings derives from their Free Will. We cannot apply the same concept in the same way to inanimate random processes. Nevertheless, there is a striking parallel between the two that can only be a starting point for a deeper understanding of the nature of both random initiative and Free Will.

The concepts that I have developed here provide insights for me personally as I consider my sister Etta Z"L. They provide insights both into her life and into the fact that she was taken from us so young.

On the one hand, her passing away is to be understood as part of the respect and love that HKB"H has for the laws that govern the Universe. The laws of physics, biology, genetics and molecular dynamics must in most cases be allowed to run their course. Until cures are developed, diseases will kill their victims. The best people together with the less good people will die. G-d may sometimes intervene here and change things there, but only well within random variations on the scale that we can see. Etta's passing was not some part of a plan. Her particular death did not serve some higher good.

On the other hand, Etta's life embodied exactly what G-d expects from humanity. Her life overflowed with initiative and drive towards the goals of a better Universe as she understood them. May her memory be blessed.

1 Neural capabilities are not, in fact, rigidly represented neuron by neuron the way computer programs require are built. Neuron states are also not binary the way coins are either heads or tails. I am using Neurons here to represent some state of possibility in the configuration of the innate capability of the brain. Any state can be theoretically represented as a sequence of 1's and 0's.