Halakha and the Laws of Robotics (Part 1 of 3)

In the course of this three part article series, we will examine the fundamental nature of the halachic system. To do this we will begin with the “I Robot” stories written by Issac Asimov.
In these stories, the three “laws of Robotics” which all robots are designed to follow are:

1.    A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2.    A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.

3.    A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

Presumably, these laws are designed to make robots safe and effective. However, strict adherence to these laws can sometimes have undesirable consequences. For example, in the story Little Lost Robot, some robots’ first law was modified since, under the regular law, the robots would have prevented humans from doing research with a type of radiation which, although dangerous in the long term, was not particularly damaging to humans in brief exposure. In another story, Liar, a robot lies to humans since telling them sensitive information would have hurt people’s feelings, thus harming them and violating the first law. But in the long term, this behavior ends up causing more harm than good. There are many other examples in the stories of the laws being less than ideal. We may wonder then, is there any way these three laws can be made ideal. For example suppose there were just one law of robotics:

1. A robot can only do the kinds of things most people would support.

This law would certainly prevent robots from harming humans in the vast majority of circumstances but would allow robots to harm humans under the right circumstances, such as the examples mentioned above. Wouldn’t this law yield robots yield better results than the three mentioned above? Why isn’t this the only law of robotics in Asimov’s stories?

The obvious answer (aside from the fact that it would make for less interesting stories) is that even if you assume that the rule has an unambiguous meaning, the meaning would probably be something that robots cannot evaluate. We might be able to teach a robot what constitutes harm, but even in science fiction we cannot teach a robot to correctly evaluate what its manufacturers would say about the merits of any particular action. So instead of the one law, we have the three. In the vast majority of circumstances, these three laws will capture the “kinds of things that the robot manufacturers want robots to do.” But creating a set of laws understandable to robots  which captures that in all circumstances is impossible.

To state the previous point in more general terms, the laws of robotics form a code of conduct which aims to approximate a different, more ideal code of conduct. For any code of conduct, it’s reasonable to ask if it itself is ideal or if it merely aims to approximate a different, less implementable code. Codes which purport to be moral codes, for example, are not likely to be approximations. The value system which asserts that people should always do whatever leads to the most good for the most people is probably not an approximation for some other, more inherently valuable code. Legal codes, on the other hand, are less likely to be ideal and instead exist because they make life livable and are relatively easy to enforce.

It is reasonable to ask which of these two categories of codes of conduct includes the Halakhic system. On the one hand, the Halakhic system is a legal system designed for human society, so it must be comprehensible to humans even if every detail of the Divine Will is not comprehensible to humans. On the other hand, the Torah is a set of laws which was communicated by God himself. Surely an omnipotent deity would be able to perfectly capture His will in His Torah. As the Ran says in his drashot, ‘The torah does not prevent good actions but rather commands their performance’ (Drasha 10). Taken to its fullest extent, this Ran would seem to imply the Torah’s commands are identical to the ideal of goodness and thus to The Divine Will. In order to determine whether Halakha is identical with or merely approximates the Divine Will, we will look at both the writings of Jewish scholars and the structure of Halakha itself.

First we will examine what others have said (or implied) about our question. The Sefer HaHinukh begins with the first mitsvah in the Torah, Peru U’rivu. Now, why were we commanded to engage in procreation? If Halakha is identical to the Divine Will, then we must assume that the book begins with the commandment to procreate because the Divine Will requires humankind to engage in procreation. However, the Sefer HaHinukh explains the reason for this mitsvah as “so that the world should be settled, since God wants the world to be settled.”

Procreating is not identical to “causing the world to be settled.” A doctor who delivers babies is causing the world to be settled but is not engaged in procreation. There are imaginable scenarios in which the process of getting married would require someone to spend a lot of time dating, forcing him to drop out of medical school and leading to a net reduction in how “settled” the world is. True, in the majority of circumstances, procreating will lead to the world being more “settled,” just like in the majority of circumstances we don’t want robots to allow humans to come to harm. But there seemingly are real cases, according to the Sefer HaHinukh, were Halakha and the Divine Will fail to converge. Why, then, didn’t God just command us to “cause the world to be settled?”

That formulation, though accurate, would be too vague. In what way should we cause it to be settled? To what extent? Since Halakha needs to function as a legal system, it needs to be precise, and sometimes, for the sake of precision, conveying the ideal message must be compromised.

This phenomena is not unique to Peru U’rivu, but rather exists throughout the Torah. For every mitsvah, the the Sefer HaHinukh gives a reason why the mitsvah was commanded. God prohibited the gid hanashe, the sciatic nerve, so the Jewish people will remember to trust that their enemies will not destroy them. God commanded the sacrifice of the paschal lamb as a remembrance of the miracles that God performed for the Jewish people in Egypt. We can’t walk too far beyond the city limits on Sabbath so that we will remember that the world has only existed for a finite amount of time. In all cases, God commanded a mitsvah in order to achieve a given ends. A mitsvah is not an ends in and of itself.

The Sefer HaHinukh was not the only one to perceive of mitsvot this way. Maimonides poses the question, ‘are mitsvot purposeless or do they have rationales external to themselves?’(Guide 3:26)  He does not even consider the other option – that the mitsvot themselves are an ends on their own. He concludes that the mitsvot do in fact have reasons, even if particular details within the mitsvot might not. He says that the mitsvot overall serve to make society peaceful and to teach everyone correct opinions. Maimonidies, like the Sefer HaHinukh, does not consider any cases when observance of mitsvot, because of particular circumstances, has the effect of making society less peaceful or teaching people incorrect opinions. However, it is obvious that such cases are at the very least possible. For example, the law that kohanim get special gifts due to their lineage might lead to resentment and disunity among the Jewish people. It is also obvious, however, that the mitsvot would still be binding in those cases. Halakha is not a perfect moral code but an imperfect legal code.

Indeed, anyone who gives external reasons for the mitsvot is implicitly saying that the Halakhic system is not the ultimate ideal but is a means for achieving these external reasons. Such people include rationalists like Maimonides and the Sefer HaChinuch as well as kabbalists who say that the performance of mitsvot has mystical effects. Is there anyone at all who believes that the mitsvot themselves are an ends – that Halakha is a system devoid of external rationales?

We find such an opinion in the writings of Yeshayahu Leibowitz. In his essay, Religious Praxis, The Meaning of Halakha, he writes, “Halakhic observance as a way of life, a fixed and permanent form of human existence, precludes conversion of religion into a means to some ulterior end. Most of the mitsvot are meaningless except as expressions of worship.” Thus, according to Leibowitz, the Halachic system could not possibly be a means of achieving the end of approximating a more ideal system, since the mitsvot are not a “means” at all. Regarding the reasons given for the mitsvot which seem to imply that the mitsvot are a means, Lebowitz says, “The rationale of a mitsvah is service of God, not a utilitarian interest.” While this could be understood to imply that the mitsvot are a means to achieve the goal of serving God, this end is identical to the observance of the mitsvot themselves, and thus Halakha is not an approximation of some external value system.

We will conduct an independent examination of the Halakhic code in part 2.