Dynamics and Statics in Halakha and Negative Attributes: The Case of Giving a Get (Column 707)

With God’s help.

Disclaimer: This post was translated from Hebrew using AI (ChatGPT 5 Thinking), so there may be inaccuracies or nuances lost. If something seems unclear, please refer to the Hebrew original or contact us for clarification.

In the two columns before the last, I began from a work by the painter Sam Back and discussed the complex relations between dynamism and statics. We saw there is no equivalence between the two planes. Sometimes one represents the other; sometimes one is real and the other subjective, and sometimes the opposite. We noted that in films, frames are what exist in reality and the dynamism arises from them only in our minds. By contrast, with a real arrow in flight the situation is reversed: in reality there is a dynamic process, but our cognition represents it by static frames (and the brain reconstitutes that into a dynamic event). In the last column I tied this to different aspects of the time axis, to the uncertainty principle in quantum theory, and to additional topics.

In the previous column I mentioned there are Talmudic examples where this distinction can be applied. The most prominent is the wide Talmudic discussion about delivering a get (bill of divorce) (see a brief treatment here). In this column I wish to address it and the broader philosophical and logical ramifications tied to it.

Giving a Get

There is a verse in Deuteronomy (24:1) that addresses divorce:

“When a man takes a woman and marries her, and it comes to pass that she finds no favor in his eyes because he has found in her some indecency, he shall write for her a bill of severance, place it in her hand, and send her from his house.”

The closing clause—repeated again in verse 3—may be the most expounded verse in the Oral Torah. Every word in this snippet of the verse serves as a basis for numerous complex sugyot, especially in tractate Gittin. Our concern here is with the words “and he shall place it in her hand,” from which we learn that divorce is effected by delivering the get (the bill of severance) from the husband to the woman.

The Talmudic question is: what exactly is that delivery? What action must occur for the divorce to be valid? At first glance, it seems simple: the physical transfer of the written get from the husband’s hand to the woman’s hand. But when one examines the sugyot in Gittin, it turns out the matter is not so simple. There are many dozens of distinct sugyot that try to get to the bottom of this delivery: a get in her hand while its cord is in his hand (the get is tied by a string held by the husband); “take your get from the ground”; conveying a field within which the get lies; delivering the get to the woman’s courtyard; writing it on the hand of her maidservant (asleep? bound?); throwing the get in public property or in the woman’s domain; into her hand, into her courtyard, or simply onto the ground; conveying the get and the courtyard together to the woman; delivering it while attached; writing it while attached and then detaching it; and on and on. It truly doesn’t end.

Naturally, commentators try to take all these sugyot and extract from them a general definition of a valid act of delivery. The Rishonim do so, tersely and succinctly as is their way; the Acharonim do so at great length and in a more analytic fashion. In a moment I will describe something of these heroic efforts, but already here I will say: they all fail miserably. If you examine the matter, you will find the definitions reached at the end are extremely intricate and, ultimately, rather vague. The number of details and branches in these definitions is roughly the number of examples on which they are built. I must clarify a bit more what is problematic about such a definition.

Two Types of Scientific Explanation: Explaining by the Unknown

What the commentators are doing in such sugyot is a kind of scientific work. In science you observe many occurrences and try to extract from them a general principle—a theory—of which they are all particular instances (this is Carl Hempel’s deductive–nomological schema; see columns 506 and 537). Thus Newton observes different phenomena—massive bodies falling to earth, planetary orbits, tides—and extracts a single general theory, gravitation, by which all these are explained, i.e., of which they are particular cases.

Note there is a step here that on its face seems quite odd. We are used to explanations based on “reducing to the known,” i.e., explaining unfamiliar phenomena in terms of known principles. For example, an accident investigation committee probing a plane crash tries to find an explanation for that unknown event in terms of aeronautical principles. They might discover a crack in a wing that could not withstand the stresses required. In that way we explain an unfamiliar phenomenon in terms of familiar, known principles. But Newton took what seems the opposite move: he took a set of familiar, known examples and explained them in terms of an unfamiliar principle he had just formulated (the law of gravitation). In what sense is that an explanation? Why are things more understood now? This can be understood through a distinction proposed by the sociologist and philosopher of science Thomas Kuhn (see column 647).

We often think a scientific theory is one that has been proven. But, as is well known, Karl Popper showed that a scientific theory cannot be proven. For example, the theory that all ravens are black cannot be proven scientifically, because to prove it we would have to observe all ravens that exist and will exist—impossible. Refuting the theory, however, is in principle easy: one raven of a different color suffices. Popper thus proposed that a scientific theory is one that is empirically falsifiable—i.e., it has predictions that can be tested in the lab and might fail (the prediction might not hold). Popper assumed that once we reach a single counterexample we must abandon the theory. It has been empirically falsified.

But Thomas Kuhn showed that in practice science does not work like that. If there is a theory that explains many facts and works well, the scientific community will not abandon it so quickly even if there are experiments that refute it. Scientists will wait for explanations that incorporate those facts into the theory or show the experiment was flawed. Kuhn attributes this to the community’s conservatism; here I disagree—this is how it ought to be, for scientific (not sociological) reasons. In any case, what matters for us is Kuhn’s description.

Kuhn distinguishes two stages in the activity of scientific communities: first, working within the paradigm, until a crisis requires replacing it. So long as there is a well-functioning theory—a scientific paradigm—the main work done is to explain various phenomena within it. Scientists examine case after case, test the theory through different experiments, yet all within the existing paradigm. From time to time an experiment arrives whose results do not fit the theory’s predictions; the community puts it aside with a cautionary note. When enough such experiments/facts accumulate, a paradigmatic crisis forms. The community concludes the paradigm has failed and a new one must be sought. This is the second type of scientific activity: now the community seeks an alternative theory that can explain all known cases; once found, it becomes the new paradigm. The cycle repeats: work within the paradigm; if a crisis forms, seek another; and so on. Thus science progresses.

Notice that the emergence of new scientific theories always follows crisis. Work within the existing paradigm expands knowledge but does not advance us to broader understanding. It does not examine the paradigmatic principles themselves but only uses them. When working within the paradigm, explanations reduce to the known, explaining more and more cases via the existing paradigm. In a crisis, by contrast, we seek a new theory (as with Newton and gravitation), and there it is clearly a process of explaining by the unknown: we take known cases and search for a not-yet-known theory to explain them.

Why is Explaining by the Unknown an Explanation at All?

At first blush, explaining by the unknown is not an explanation; it’s like a Hebrew–English dictionary that takes known Hebrew words and “explains” them with unknown English words. Still, we all grasp this is a different kind of explanation. The reason is that this process illuminates the common foundations of many cases and facts that appear different on the surface and shows they all obey a single general principle (or a set of principles—a theory). We see they are all particular instances of a general theory. Even though the theory was just formulated and is unfamiliar, it has explanatory power and clarifies reality, because it simplifies it and reduces all these events and situations to a small set of principles.

Now ask: what would you say of a theory whose number of principles equals the number of examples it purports to explain? Such a theory has no illuminating power and seems pointless. Instead of listing 100 facts, we describe them as particular cases of a theory that includes 100 components/principles. We’ve gained nothing. Mathematics makes this vivid. Consider the sequence …1,3,5,7,9. One may “explain” it by saying it’s the sequence of odd numbers—one principle that covers infinitely many members, hence explanatory. Now consider …1,3,8,19,20,30,33,37. Here’s a possible “explanation”: start at 1, then add 2 (get 3), then add 5 (get 8), then add 11 (get 19), then add 1, then 10, then 3, then 4, etc. Is that satisfactory? Of course not. It’s merely a re-description that yields no insight. It brought no order to the sequence’s structure and gave us no small set of principles that “explains” it. The tell is that the number of components in this “theory” equals the number of terms. Such “explanations” are usually worthless.

What Is Giving a Get?

When defining the process of giving a get, commentators survey the varied examples in the Talmud and try to extract a “theory,” i.e., a general definition of “giving a get,” of which all the cases in the Talmud are instances. As noted, this fully parallels scientific work as described above. For those interested in detail, the most systematic, detailed, and comprehensive source I know is Kehillot Yaakov (KY) to Gittin, §§14–17. These four sections have different titles, but all try to define “giving.” Try to follow and you’ll see it’s sheer madness. He marshals many examples (still far from covering everything in the Talmud), and the proposed definition, as the discussion advances, becomes more and more obscure—sounding increasingly ad hoc. In effect there are de facto fits to the Talmudic cases, and I arrive at a strong feeling of overfitting (see the problem in columns 243 and 426). To give you a taste, I’ll sketch a bit more—very briefly.

One can start with the simple definition that arises from the verse itself: giving is transferring from the husband’s hand to the woman’s hand. That obviously fails many examples. There are many cases where that does not occur and yet the Talmud says the divorce takes effect. For example, conveying to the woman a courtyard within which the get lies; or perhaps conveying the get itself to the woman (there are contradictions between sugyot here, prompting an extensive subtopic in the Acharonim: proprietary rights in the get; see KY §17). Fine—perhaps what matters is specifically conveyance (kinyan) of the get from husband to wife and not physical delivery. That is a more precise definition that fits many more examples. Note that physical delivery could itself be a kinyan (pulling or lifting). But the Ketzot HaChoshen (§200:5) already noted that even this does not fit all Talmudic examples. He rejects it, for example, based on the sugya “write it on items prohibited for benefit,” which the Talmud validates (see Gittin 20a), even though according to some Rishonim there is no ownership in such items; i.e., there was no conveyance and yet the get is valid. So it is neither of the two simplest options—neither physical delivery nor conveyance. What, then, is the “giving” the Torah speaks of?

From here on, KY tries to move toward a more complex definition that will fit all the examples, but the matter becomes more and more convoluted, the formulations more and more obscure, and they sound increasingly ad hoc. It actually seems like case-by-case fitting to the Talmud, and we get the strong sense of overfitting. To give you a sense, here’s a bit more of the process.

At the beginning of §14 he discusses the sugya “Here is your get, but the paper is mine” (Gittin 20), about which the Talmud rules she is not divorced. Apparently, then, conveyance is required—contrary to our earlier conclusion. Rashi explains that indeed conveyance is not required, but “giving letters” is not giving; i.e., the problem is not the lack of conveyance but the failure of a minimal requirement of giving. But what is that giving? Not clear. He then discusses one who “retains” the parchment between the lines (gives only the parchment under the letters, not the parts between the lines; here he gives her something, yet in some sense the letters still hang in the air). Then he discusses one who wrote the get on a slave’s hand or on a cow’s horn and gave her only the horn or only the slave’s hand—or both. Then he brings cases in which the husband has some physical or other “tie” to the get delivered, and there too the problem is that this is not giving (or not “severance”). He also discusses cases where she writes or he writes on paper that is hers (which may be invalid for all deeds, unrelated to “giving”).

In §15 he treats the case where the husband gives the get in his courtyard, and then writes her a deed of gift for the courtyard with the get inside it. He raises a difficulty from a messenger of delivery who became a messenger of receipt (there was no “return” of agency to the husband); there too there is apparently no giving (or perhaps a different defect). Likewise when the woman herself becomes a delivery agent, which is analogous. This may also depend on who performs the kinyan on the courtyard (“she locked and opened”), and perhaps it resembles “take your get from the ground” (note there is a conveyance of the get from him to her, yet it is ineffective because there is no act of delivery). Then he moves to “he loosened his belt and she pulled it” (she draws the get from his pocket, depending on whether he moves his body, assisting her in taking the get).

Here is a mid-paragraph excerpt from §15, just to give a feel:

“Certainly conveying ownership of the get itself does not suffice to be considered ‘giving,’ for besides conveying the get we require that the husband assist in the kinyan-act of receiving the get. But conveying the courtyard, which serves as a ‘hand’ for receiving the get, is indeed considered his assistance in the kinyan-act of receiving the get, since without his conveyance there would be no kinyan-act here, and it is through his conveyance that a kinyan ḥaẓer (courtyard) of the woman is effected. In every case of ‘take your get from the ground,’ without the husband’s conveyance the woman herself performs the kinyan-act (since it is in her hand); only for acquisition (zakhiyyah) does she need his conveyance. But when the husband conveys the courtyard in which the get lies, it is only through his conveyance that the woman’s kinyan-act comes about.”

This mess is a summary of the first part of the second of the four sections dealing with the issue. Note how many components we have swept into the definition: roughly as many as the number of examples that generated it. If you try to summarize the general definition that would emerge from all four sections, its complexity would rival that of the sections themselves. So much for the description; I hope it suffices to illustrate the difficulty.

The Difficulty: Another Look at Casuistry

The fundamental question is: why didn’t the Talmud simply give a general definition? Why did it discuss dozens and hundreds of cases—augmented by generations of commentators—so that we would try (unsuccessfully, as we saw) to extract a general definition from the mass? If we assume the Sages had some definition—otherwise, how did they rule in the myriad cases they discuss?!—why not give it to us directly?

This can be answered in terms of casuistry. In column 482 I discussed the Talmud’s casuistic character and its advantages over a positivist approach. The Talmud prefers to reason through cases and examples and leave it to commentators and decisors to make analogies from them, rather than engage in analytical, conceptual discussion through general principles and deductive derivations (a positivist method). I explained this stems from the complexity of legal/halakhic topics, which does not permit exhaustive analytical discussion. On the contrary, clinging to general principles often leads to wrong conclusions because we miss shades and nuances between examples and the limits of deduction.

I already mentioned there the sugya of giving a get; now we can see it more explicitly. It seems to me the Sages had no explicit definition of “giving a get.” They formed their positions on the myriad Talmudic cases intuitively, not by deducing from some general principle. That is why they did not provide us such a definition: they didn’t have one. They try to convey to us their intuition by means of many labeled examples (“valid”/“invalid”). Below I’ll note this is very much like training a neural network.

In the case of giving a get there is a specific difficulty in offering a general definition. We saw the Talmud and commentators examine cases described by their endpoints—what was before the giving and what after. Thus, for example, both delivery and conveyance are described via endpoint states. But the act of giving is dynamic, and as we saw in the two columns before the last, one cannot capture the essence of dynamism from the static endpoints through which it passes (as with understanding speed only through positions at different times). Precisely because we are trying to define a dynamic act, an attempt to define it through endpoints is doomed to fail. This is a special complication in the sugya of giving a get, though in most other topics the Talmud likewise prefers casuistry over general principles and deductions. Dynamic concepts simply provide a particularly apt illustration.

The right way to handle complex concepts, and in particular a dynamic event, is through examples. Examples allow us to grasp an abstract that resists direct definition by looking at cases and trying to infer from them what will meet the requirements of giving and what will not. After encountering many examples, we develop insights and our intuition can judge in each case whether there was proper giving or not. This closely parallels what we saw in columns 694–699 regarding training a neural network. There, too, instead of giving a general definition as in classical software, we prefer to train the network by labeled examples (“right”/“wrong”) so that it develops an internal structure that can handle future cases. Training by examples bypasses the need for a general, explicit definition—very hard to attain in complex cases. A review of that series can illuminate the claims here. Think of wanting to teach the network to recognize dynamic occurrences. We have no way to define “dynamism” explicitly (only via successive frames). So if we want it to detect dynamism itself, we must feed it varied examples, as I did in previous columns and as the Talmud does regarding giving a get. That is exactly what the Talmud tries to do for its learners.

Reflections on Yeshiva-Style Learning

Given all the above, why do we, the learners and commentators of the Talmud, try to conduct analytic discussion and extract from the Talmudic examples a general definition? Note that this is what is usually done in conceptual yeshiva learning. We examine the examples and try to extract a general definition that captures them in a comprehensive picture. This is conceptual, analytic, overtly positivist—quite the opposite of Talmudic casuistry. Regarding giving a get, we saw here that this attempt did not really succeed. Why engage in it if the Talmud itself teaches us the attempt is futile (except in very simple cases)?

Let me broaden the lens. Anyone with experience in analytic yeshiva study knows how limited it is. We always begin with two polar “chakirot” (investigative alternatives), but almost never does one side suffice to explain the sugya in all its complexity. We always continue complicating the analysis, blending the sides, adding exceptions and branches to the simple initial dichotomies—just as we saw in KY here (see my article here and elsewhere). Why do this at all? Do we hope to succeed where the Talmud itself “despaired” from the outset? Such hope is not illegitimate, of course. I do not think our failure is guaranteed where the Talmud did not supply a definition. But the fact is that in many cases it doesn’t really work. So why insist on this mode of study? Rav Ovadia indeed looked quite critically at the yeshiva methods prevalent in Ashkenazic institutions, viewing them as (unaspirated-peh) pilpul—empty contrivances.

The yeshiva style merely sharpens the difficulty; yet this is true of commentators throughout the generations. All engaged in analyses and attempts to define concepts and draw conclusions analytically. Is there any point?

A Resolution of Analytic Study

I think the way to understand the meaning of analytic learning is to follow KY’s flow in those sections. The Sages indeed worked intuitively and ruled on the myriad Talmudic cases. We then seek a general definition of “giving a get.” First we propose the hypothesis that giving a get is its physical transfer from the husband’s hand to the woman’s hand—the simple reading of the verse. Immediately we see this fails. We then propose that giving a get is conveyance (kinyan). Very quickly that fails as well. Are we now back where we started? Not at all. We have learned something about “giving a get.” Not only that it is neither transfer nor conveyance, but that there is something else present that is sometimes realized by transfer and sometimes by conveyance—though neither is necessary. We have learned that both are examples of a more general definition, two particular cases of it. We keep moving, proposing more complex definitions and again rejecting them—and again the rejection teaches us something. We build within ourselves a better and better grasp of the concept of giving, even if we have no explicit verbal definition. Each rejected possibility adds another layer to our nonverbal understanding and sharpens it, bringing us closer to grasping what “giving a get” is. The point is this: the understanding formed at the end is not exhausted by the final formulation, because sometimes there is no such formulation; sometimes, even if there is, it is not worth much—at least if we treat it as a general law. The formulation merely helps us reject further possibilities, repeatedly honing our intuition.

We are built such that our analytic thinking proceeds analytically. Unarticulated intuitions do not advance us. We must formulate hypotheses and try to corroborate or refute them, then refine them, and so on. The ultimate goal is not achieved only when we have a formulation that covers all examples. Even if we do not reach such a formulation—or if it is so complex that it is of little theoretical value—it hides behind it the entire analytic path we have trodden. In short, it is the path that has built our core understanding, not necessarily the final formulation. Return to the excerpt I quoted above from KY and you’ll see it indeed adds understanding, even if it does not provide a positivist, logical tool from which we can deduce answers for every future case. That is not its role. It merely summarizes where we currently stand and the ensemble of understanding we have amassed by all the negations along the way.

This returns us to the earlier parts of the column. You can see how similar this is to scientific practice. Popper taught that a scientific theory cannot be proven, only refuted. The role of counterexamples in scientific experiments—or of analytic cases in a sugya—is just that. We propose an initial hypothesis, test it against cases, and it undergoes refinement (articulation). There, too, the negations build an intuitive understanding within us, and it is that—rather than any ultimate formulation—that constitutes the real product of the scientific/analytic process. As we saw, this also resembles a neural network, trained by examples (some confirming, some refuting the current state—i.e., the hypothesis under test or the existing paradigm), which sharpen and hone its internal structure so that an intuitive “understanding” forms—one that does not rest on articulated rules (as in classical software). Unlike humans, a network has no real understanding, whereas in us that honing is accompanied by a mental dimension we call “understanding.” Still, the idea and process are very similar.

Hence it should not surprise us that Brisker-style dichotomous formulations don’t really work and don’t truly explain all cases. Life is too complex to be forced into such an analytic mold. But because our thinking works analytically, we proceed by formulating sharp theses or hypotheses and then confirming or refuting them through cases. The analytic “chakirah” is a tool for sharpening intuition; the aim is not necessarily to reach a final, exhaustive formulation.

In a certain sense, the analytic path that characterizes our thinking is analogous to a static perspective (through an ideal camera, in the terms of column 705), whereas intuitive (synthetic) understanding is analogous to direct cognition of dynamism (obtained through an ideal movie camera). There, too, we saw we are bound by our modes of thought and cognition, through which we strive to understand and describe reality. What’s important is to be aware that these are models arising from our cognitive modes, not reality itself. We saw there that the derivative does not define speed; it only describes it in the language of our cognition (which is static, hence limited and unable to comprehend dynamism at an instant). So, too, our analytic thought is only a model of halakhic and scientific truth; we employ it to understand reality only because our thinking and cognition operate in static frames. But we must internalize that this is not reality itself, and therefore we should try to rise above it to form an understanding of reality itself. Analytic tools remove sharp, dichotomous possibilities and thus build within us a more “rounded” intuitive understanding (on intuition and synthetic thought see, e.g., column 653, and in greater detail in my books Two Wagons and Truth and Not Steadfast).

This brings me to an interesting philosophical point that can be clarified from our discussion: the doctrine of negative attributes. I’ll touch it briefly here only to clarify its linkage to our topic.

The Doctrine of Negative Attributes and Its Lack of Content

It is known that Maimonides and other philosophers across religions argue that God cannot be described directly; all attributes applied to Him are necessarily negative attributes. This “negative theology” makes at least two central claims: (1) no positive attribute can be ascribed to God; (2) negative attributes nevertheless teach us something. As an aside, the kabbalists claim God can be described with positive attributes (I will not go into exactly what that means or their relation to His essence), and this is a key dispute between philosophers of religion and kabbalists.

The fundamental question regarding this theology is logical. If a negative attribute teaches us something, then how does it differ from a positive attribute? Why may we use it of God if we may not use a positive one? For example, suppose I say God is merciful. According to negative theology, I have said He is not cruel (He is “not non-merciful”). But double negation cancels out and we are left with “He is merciful.” What have we gained by moving to a negative description?

One might explain the problem differently. Aristotle already distinguished two types of negation: contrary negation moves me from +1 to –1, while privative negation moves me from +1 to 0. In reality: the relation between cold and heat is contrary negation, for if we combine heat and cold (turn on heater and AC, like here—at the end) they cancel each other (like +1 and –1), yielding a lukewarm temperature. By contrast, the relation between light and darkness is privative: combining light and darkness yields light; there is no cancellation. Thus if light is 1, darkness is 0. On that basis, one could say that negative attributes are privative, not contrary; i.e., when we say God is merciful we mean He is not cruel in a privative sense: cruelty does not apply to Him. This means the mercy–cruelty axis we use for humans is inapplicable to Him. But by that logic, we could also say He is cruel, for that would be interpreted as “He is not merciful” (privatively), i.e., that the mercy–cruelty axis is inapplicable. If so, we could ascribe to God any attribute whatsoever; in any case it would say nothing about Him except that nothing can be said. Why, then, does Scripture give thirteen attributes describing Him? We could equally have used the opposite attributes—or just a single “attribute” stating that no attributes apply.

In short, negative theology seems like contentless nonsense. Either we describe God or we do not; wordplay with types of negation leads nowhere else. This difficulty has bothered me for many years regarding negative attributes.

A Possible Resolution: ‘Ḥak Tochot’ Explanations

In light of the above, we may propose a resolution of this enigmatic doctrine. We saw that analytic study of halakhic concepts/principles proceeds in a way that can be seen as a chain of negations. We negate from “giving a get” various simple, sharp descriptions—physical transfer, conveyance, etc.—but, as we saw, these negations produce in us an abstract understanding of the concept even though it resists explicit verbal definition. The analytic path serves to reject simplistic understandings that fail the halakhic facts, but the series of negations yields a nonverbal yet positive understanding of the concept—much like training a neural network.

This can be likened to the halakhic notion of “chak tochot” (see e.g., Gittin 20a). The Gemara there discusses what counts as “writing” a get and, by extension, what constitutes the melakha of writing on Shabbat. It considers one who writes letters passively by carving out the background around them. For example, a person forms the letter kaf by digging out the inside and outside around the letter so that the shape of the letter remains raised. In such a method one never touches the letter itself, only its surroundings, yet in the end the letter is formed. The act is one of removal only, but an inscription results. This is similar to a sculptor chiseling a statue from a block of stone or wood. Seemingly the sculptor does nothing but remove what is not the statue; nevertheless, that is how a statue is made. Likewise, sometimes analytic study merely removes dross (i.e., errors and partial definitions), but in the end—via “chak tochot”—we are left with a positive, intuitive (usually nonverbal) grasp of the concept or principle under discussion.

I would suggest that perhaps proponents of negative theology intend negation of this kind. When we say that God is merciful, we mean He is not cruel in a privative sense, and by that we carve away cruelty from our conception of Him. What remains is not “mercy” in the human sense (as would be the case with contrary negation), but something analogous that must be abstracted when applied to God. Thus we could not equally say He is cruel, for the intent is not that the mercy–cruelty axis altogether fails to apply; rather, the human description of those traits does not apply. Yet something of the essence of mercy (not of cruelty) does pertain to Him—at least in the sense that cruelty is excluded in every way, even in abstracted form. By contrast, saying He is merciful is right of Him, albeit in an abstracted way. He does not feel emotions of mercy as we do (He does not have sensation as we do), yet saying “He is merciful” is the best description we have. [1] This is why Scripture offers true descriptions (attributes) of God, and uses these rather than their opposites, as I wondered above.

I do not know if this is what the advocates of negative theology intended, but it is the only direction I can think of that lends some meaning to this puzzling doctrine. All we do in analytic study of halakhic concepts/principles is, by “chak tochot,” develop an intuitive grasp of them. In that sense, lomdus resembles negative theology, only it treats halakhic topics rather than descriptions of God. This is also how we understand dynamic processes: it is a nonverbal yet decidedly positive understanding, obtained by clearing away the static frames to glimpse what underlies or lies beyond them.

[1] One could interpret this only in terms of attributes of action, i.e., that the statement “He is merciful” pertains only to His actions, not to Him. The claim would be that His actions look like ours when we act mercifully. But on that reading we are speaking of attributes of action, whereas the doctrine of negative attributes concerns attributes of essence (many negationist thinkers hold that attributes of action do not require any negative approach; they may be ascribed positively). Thus my claim here is that saying “God is merciful” pertains to Him, not only to His actions; but what is ascribed is not mercy in our emotive–experiential sense, rather something abstract that in some way resembles human mercy. The actions that flow from His being merciful are indeed merciful in the sense of attributes of action, but, again, negative theology is concerned with essential attributes, not attributes of action.