Raphael Magarik

Romance and Rational Choice


Variant texts are properly thought of as frenemies. Stemmatics imagines textual families, in which deviations and change are natural results of time and chance. But we also speak of “conflicting” readings. Only one can be promoted to the main text, its rivals consigned to the apparatus. I want to look at a pair of such texts, each of which contrasts two alternate models of dyadic friendship.


The text(s) come(s) from Pirkei Avot, a collection of wise sayings included within the Mishnah, a canonical Jewish lawcode redacted roughly at the end of the second century CE. Here is the more familiar text of mAvot 5:16 (and the one found in the Kaufman manuscript):

Any love that depends on something: when the thing ceases, the love ceases.
But a love that doesn’t depend on anything never ceases.
What is a love that depends on something? The love of Amnon and Tamar.
And that doesn’t depend on anything? The love of David and Jonathan.

Before interpreting, I want to insist first that the incessant structural doubling is not just an artifact of later manuscript corruption. The mishnah, which discusses investigates dyadic friendship, splits them into two models and is itself split in two parts, an abstract first half and then concrete examples. Further, it forms a larger whole with its immediate neighbor and twin (mAvot 5:17), an identically structured mishnah that dichotomizes conflicts, rather than loves. The mishnah is clearly thinking about pairing in its form as well as its content.

Substantively, the mishnah contrasts what we might call “unconditional” and “conditional” loves, privileging the former. “Unconditional love” is of course a minor mystery, something which defies the Principle of Sufficient Reason. If a love doesn’t depend on something, is it uncaused? How much more is there to say about it? As the nineteenth century Mishnah commentator Rabbi Israel Lifschitz (known as the Tiferet Yisrael) writes, “When we ask this person why they love all these, they don’t know how to give the reason of the matter.”

There are, to be sure, reasons for the unreasonable. The old riddle about love that depends on something is that it seems to be love of that thing—not love of the person in whom that thing happens to inhere. That is true for lofty, spiritual “things” as much as for passing physical ones. As the Tiferet Yisrael writes, “That a person loves a wise man, or a virtuous person, or a righteous person—that’s also considered ‘dependent on something,’ because when the wise man will cease busying himself with wisdom etc., ‘when the thing ceases, the love also ceases.’” The ethical problem is something like this: if you love a sage and she develops Alzheimer’s, say, or goes to business school, it matters quite a bit whether you loved her “for her,” so to speak, or for her wisdom. The worry about the latter case is that it instrumentalizes the beloved, treating as a means a person, who ought to be regarded as an end in herself. And such instrumentalization (loving someone for who he is for me, rather than for who he is for himself) terminates logically in narcissism; The Tiferet Yisrael says, “He doesn’t love the beloved person, only the thing he will receive from him in the future, and it turns out he only loves himself.”

Nonetheless, we only have access to the phenomenal, and “unconditional love” remains necessarily indefinable, a problem which may motivate the following alternate version of Avot, attested to by, for instance, Maimonides:

Any love that depends on something frivolous [b’davar batel]:
             When the thing ceases [batel davar], the love ceases.
But a love that does not depend on a frivolous thing never ceases.
What is a love that depends on a frivolous thing? The love of Amnon and Tamar.
And that doesn’t depend on a frivolous thing? The love of David and Jonathan.

The Hebrew adjective meaning “frivolous” is spelled just like the verb meaning “ceases.” Thus, the texts differ just regarding whether the word “batel” appears once or twice in the first line. Because both removal via hypercorrection and duplication are plausible, and because, if I am right, neither philosophical option is completely satisfactory, I do not know whether Kaufman’s concision precedes Maimonides’s antimetabole or the reverse. If only because this essay argues for a certain kind of textual mutuality, I intend to treat them as equally plausible alternatives.

In the second version of mAvot5:16, only inappropriate mediating objects threaten the persistance of the love, implying that there are appropriate mediating objects. “Any love that doesn’t depend on a frivolous thing but rather on a durable thing,” explains R. Yom-Tov Lipmann Heller, author of the seventeenth century mishnah commentary Tosafot Yom Tov, “like the love of the righteous and sages, will never cease.” Maimonides fits the mishnah within an even more rigid philosophical context, equating frivolous things with “physical causes” and non-frivolous things with spiritual or intellectual objects. Physical things are definitionally contingent and thus pass away, whereas intellectual causes are necessary and eternal.

Plainly, we are free to imagine our own list of “non-frivolous” or “durable” goods: trustworthiness, wisdom, a good sense of humor, sensitivity, compatible political or religious values, whatever. Just as plainly, there is some truth in this account. We do evaluate potential love-objects (whether friends or lovers) based on reasons, and we reason about which features are likely to have lasting significance and which are not. But as I suggested above, there is something over-calculating in this account, which may lead to an unattactive erotic selfishness.


How are we supposed to relate these two versions of the mishnah? The formal doublings of the text suggest to me that the mishnah is self-glossing, that is, that we can produce a hermeneutic for reading the dispute I’ve outlined above from the text itself. How so? Well, most basically, the mishnah states clearly its relational goal, which is persistance; the privileged relationship “never ceases.” Further remarkable is that this goal is identical to the next mishnah’s goal for disputes (translation modified slightly from Sefaria):

Every argument that is for [the sake of] heaven's name, it is destined to endure.
But if it is not for [the sake of] heaven's name—it is not destined to endure.
What [is an example of an argument] for [the sake of] heaven's name?
               The argument of Hillel and Shammai.
What [is an example of an argument] not for [the sake of] heaven's name?
               The argument of Korach and all of his followers.

The enduring relationship hovers between friendship and conflict. Note how the two mishnahs chiastically play off different senses of “and.” The “ands” joining Hillel to Shammai and David to Jonathan are mutual conjunctions. The “and” between Amnon and Tamar clearly has a direction (he loves her, but not the reverse). Were I not rendering uniformly the Hebrew letter vav used as a prefix, we would want to translate, “Amnon’s love for Tamar.” And in the last case, we would have expected “the argument of Korach and Moses.” The pattern-breaking conjunction, which may originate in the mishnah’s reluctance to taint Moses with either association to Korach or non-endurance, signals Korach’s failure to create a truly relational conflict.

If the mishnah prioritizes persistance, how can we preserve the relationship between Tiferet Yisrael and Tosafot Yom Tov? Tiferet Yisrael suspects that love of particular traits is not really love of the person (and is thus menaced by narcissism). Tosafot Yom Tov thinks good love finds its foundation in virtuous, commendable traits. How can their variant readings of the mishnah co-exist?

If the relationships within the text are, in a sense, allegories for the relationships between the versions of the text, perhaps we answer this question from the materials of the text itself. The gay couple of Jonathan and David (who eulogizes his beloved, in 2 Samuel 1:26, saying: “wonderful was your love to me, passing the love of women”) suggests a resistance to synthesis. Pre-modern homosexual relationships are distinctive in being able to continue at length without producing offspring. Thus, when Jonathan says to David, “God shall be between me and thee, and between my seed and thy seed, for ever” (1 Samuel 20:42), he is talking about two distinct genealogical lines, two “seeds.” They share a pact or contract, which preserves their autonomy and difference, rather than merging together to become or generate one flesh. But homosexuality also intensifies the mishnah’s preoccupation with symmetry and suggests the core value of mutuality. At first, Jonathan holds power, nearly of life and death, because David is threatened by Saul; as Jonathan realizes, though, David is soon to be king. Even if not all heterosexual sex is rape, it is hard to imagine quite this easily reversible a relationship between a man and woman in the Bible.

My own sense is that ineffable, unconditional and rational, goal-oriented loves are mutually constituted. Because I suspect Tosafot Yom Tov is more intuitive to my readers, I will dispense with him more quickly: plainly, we guide our commitments by our values and desires. There is truth in the box-checking of dating websites. Tosafot Yom Tov critiques the assumptions that romantic love is ineffable or mysterious, radically separated from other social spheres of value; in short, he debunks Romance.

But if romance is in need of debunking, so is the over-simplified model of economic rationality on which that debunking is frequently founded. Our rational goals, after all, are in turn the product of the existential givens of our relationships: initially, our families, but subsequently all the people and communities to which we commit. This is a point the philosopher L.A. Paul recently brought out, arguing for the difficulty of using rational choice theory to guide major, preference-altering life decisions. Rational choice theory (which is hardline Tosafot Yom Tov) suggests people make decisions in accord with their systematized preferences. Paul argues that major choices, like having a child, actually reshape our characters and preferences such that it would be irrational to use our current utility schemas to make such decisions. The influence of your parents and your children, because of the shared “given-ness” of the relationships, illustrate a broader point: one’s values and preferences are produced by relationships and love. So Gary Becker, the economist who pioneered the economic study of “marriage markets,” ignores where homo economicus gets his utility functions from in the first place. He ignores as well that marriages do not just satisfy but also reshape the utility curve. It is fine to calculate, even selflishly, so long as you are planning the overthrow of your own calculations, of your own self. Sometimes David is on top, sometimes Jonathan.