“Eilu va’eilu divrei E-lokim Chaim, vehalakhah keBeis Hillel — These and those are the words of the ‘Living’ G-d, but the halakhah is like Beis Hillel.” The voice rung out from heaven that even though the law is ruled according to Beis Hillel, Beis Shammai’s position is also the word of G-d. There is some variation in halakhah between people — and yet both are right. We can’t expect every halachic decisor to take the same data and reach the same conclusion, even if no one errs. In the past, we compared different models for explaining how multiple correct answers could coexist (see Eilu vaEilu parts I and II). Here, that would keep the conversation too broad. I will instead just explore the problem from the Maharal’s perspective.

The Maharal’s position is that “divrei E-lokim Chaim — the word of the ‘Living’ G-d” is simply too rich and too complex to exist in this world. Therefore they are mapped to oversimplified models, related to Hashem’s words the way a shadow is a flattened representation of the original. And thus, different people looking at the problem from different directions will get different shadows — even though they are all accurate representations of the same thing.

To finish out the metaphor: The angle at which we look at Devar Hashem is our “derekh“, our path in how we . This derekh, just like the lamp, is determined by two things: mei’ayin basa, ule’an ata holeikh — from where do you come, and to where are you going? Where the lamp is, and the angle it points. Different people were put together differently, and can have different emphases in how they interpret the ultimate goal.

The complexity of Devar Hashem causes the illusion (to us) of paradox. It’s no more real of a paradox than the 5 blind men who argue about the nature of the elephant. The one who felt the elephant’s ear would argue an elephant is like a fan. The one who felt its leg would think it is like a tree. But it’s only because we can’t capture the full picture.

We therefore see the Torah as demanding conflicting values and duties. (Unresolvable dialectics, in R YB Soloveitchik-speak.) Depending upon which we choose to prioritize, followers of different derakhim will obtain different results. But you won’t make it to the top of the mountain if you first try this route and that that. You need a consistent plan.

Someone who changes the weights to find a desired result is no longer simplifying an Infinite Truth to fit it into this universe. Different shadows of the same object are each valid. But if you trace the shadow while changing the direction of the lighting mid-stream, you are left with a picture something that isn’t a shadow of the original. The weighting can’t simply be to justify the result; and in that sense even including human cost is different than ends-driven decision making (picking the pesaq to fit some non-Torah desire). The weighting system, the angle of the light, is the a priori — and must itself be a product of the halakhos of making halakhah.
This notion, that halakhah is a human-sized model of something far richer, also dovetails well with another idea I fell in love with, something from Professor Moshe Koppel’s book, “Metahalakhah”. There are two ways to learn a language: The native speaker doesn’t learn rules of grammar before using them, he just knows what “sounds right”. In contrast, an immigrant builds his sentences by using formalized rules, learning such terms as “past imperfect” and memorizing the forms that fit each category. R’ Koppel notes that the rules can never perfectly capture the full right vs wrong. A poet has to know when one can take license.

He argues that halakhah is similarly best transmitted by creating “native speakers”. It is only due to loss of our progressive loss of the Sinai culture with each generation that we need to rely on transmitting codified rules. (RMK notes in a footnote the connection between this idea and some ideas in R’ Dr Haym Soloveitchik’s essay “Rupture and Reconstruction“, Tradition, Summer 1994.) Earlier cited cases are the loss of culture that occurred with Moshe Rabbeinu’s death, when 300 halakhos were forgotten, and Osniel ben Kenaz reestablished them — chazar veyasdum. Similarly the reestablishment of numerous dinim by Anshei Keneses haGedolah after the return from the Babylonian exile — shakhechum vechazar veyasdum. Leyaseid, he suggests, is this codification.The informal knowledge of a “native speaker” is limited by the capacity of the human mind. But still, it captures more of the ineffable whole, the true “divrei E-lokim Chaim” than can be set down as formal rules.

Even the codified rules, therefore, are not all-or-nothing absolutes. Rather, they give clarity to the issues that a poseiq must weight, highlighting the relevant aspects to different elements of the fuller picture. For example, “acharei rabim lehatos” means that rulings follow the majority. But this doesn’t mean the minority is entirely ignorable; and in fact one may need to rely on that minority opinion if other factors come into play. The poseiq can then weigh pro vs. con.

We’re dealing with a fuzzy system, which acknowledges a realm where answers may be more or less halachic, rather than entirely within or outside the fold. Of course in many situations, the answer is clear and the difference between this kind of system and a straight rule-based algorithm is moot. But those are not the scenarios that require complex pesaqim and become the “interesting cases” in our responsa.

This weight-based methodology doesn’t make for looser requirements. In fact, often quite the opposite. If all rules were absolute, always of the form that factor X always trumps factor Y, then our language for dealing with conflicts in priorities would be quite limited. In cases where conflicts come into play, where the Maharal’s notion of simplification of the Infinite is manifest, one is given no guidelines — and thus full autonomy would have been granted. Rather than an algorithmic interpretation of halakhah being more defining, it’s less.

To explain further by metaphor:
If you have a digital thermostat, it probably is based on Fuzzy Logic. I’m not sure Zadeh’s “Fuzzy Logic” is the best way to represent more or less vs all or nothing, but it was that meaning that I was suggesting. There are other multi-valued logics. Statistics could even be adapted as one. Quantum Mechanics suggests a third, etc… But it is a good example to serve as a parallel for our purposes.

Fuzzy Logic is one in which AND means “take the minimum”, and OR means “take the maximum”. Say two balls are different shades of red, one a real primary red — we’ll say it’s .9 red, and another somewhat muddier, some might even call it brown — just a .2 red. In FL, we would say that the statement “both balls are red” is also a .2 (AND — the minimum of .2 and .9), while the statement “at least one of these balls is red” is a .9 (this ball is red OR that one [or both] — the maximum).

Fuzzy Logic is used in thermostats because the question “is it hot?” really needs to be able to represent “no”, “a little”, “very”, etc… Without such gradations, digital thermostats tend to turn the heat on and off too often (or need some even more complicated solution). Now, one thermostat manufacturer may weigh the heat based on degrees above comfort zone. Another might acknowledge that these things are non-linear, that I’m not nearly half as uncomfortable when the temp is 5 deg off than when it’s 10, and may have some fancier weighting system.

Notice that there are basic rules that any thermostat must comply to. For example, the temperature should be within the desired range at all times. All thermostats will end up sharing certain properties of the rules and the weightings used.

In Artificial Intelligence software, the word “heuristic” is used to describe a system for finding a solution that is less formal than an algorithm, might not always get the optimal solution, but is used because the perfect algorithm either doesn’t exist or would be too slow or complicated to ever get used. In truth, computers are algorithm machines, and thus the heuristic is really a just a much faster or simpler algorithm than one aimed directly at solving the problem perfectly.

Heuristics better represent human thought, as people aren’t algorithm machines. Heuristics better capture that looseness. Like in this case, we weigh pros and cons, not follow strict “IF … THEN …” rules on true vs false prepositions.

And so we find a range of dependence on formal rules in teshuvos. There are those I would call da’as Torah teshuvos, where it’s clear the poseiq was enough of a “native speaker” to realize the issues and weigh them before he was even conscious of needing to. The teshuvah then becomes an excercise in explaining the ruling post facto, perhaps to confirm than the instinctive answer was sound.

My favorite example of the da’as Torah pesaq is the prohibition of electricity on Shabbos. There is far more universal agreement that electricity just doesn’t fit the feel of Shabbos, as shaped by hours of talmud Torah, than on the reason why it’s assur. And in fact, some of the reasons found are self-evidently stretches to explain after the fact what seemed obvious. Boneh (building) a circuit? Where else do we see something called boneh on Shabbos where there is no actual roof involved?

Then there are the teshuvos where the poseiq relies very heavily on rules and precedent.

In between are the typical teshuvos, ones where the heuristic nature stands out quite clearly. A common form of teshuvah in Even haEzer and Yoreh Dei’ah (as well as some parts of Orakh Chaim, e.g. eiruvin) is where the author starts out by proving the law in question is rabbinic, or that some major issue doesn’t apply, and then provide snifim lehaqeil, motivations for leniency, where no one is sufficient but the poseiq combines them as justification.

---

In a future entry, I will explore how this notion is applied differently by various poseqim, and what it means to understanding machloqes and the limits of pesaq. Defining a limit is more difficult when one can’t rely on checking compliance to specific rules. How does one differentiate between differences in prioritization of issues and outright violation of the system?