Ants, Engineering, and Epistemology

Scientific American reports:

Tiny Robots Mimic Termites’ Ability to Build without a Leader

Using sensors coupled with a simple set of rules, the robots worked independently to build structures

Termites can create mounds that are hundreds of times their own size, working independently without communication or a leader. Inspired by the creatures, scientists have created robots that use just a few simple rules and environmental cues to build castle-like structures and pyramids.

The robots all work independently. Each travels along a grid and can move, climb a step and lift and put down bricks. And they use sensors to detect other robots and existing bricks, and react to these stimuli according to a simple set of rules, such as when to lay a brick or climb a step higher. The template for each three-dimensional structure is translated into a specific set of ‘traffic rules’ and combined with fixed laws of robot behavior, says co-author Justin Werfel, a computer scientist at Harvard University in Cambridge, Massachusetts. His team’s results appear today in Science.

The idea of combining traffic rules and robot behavior is “brilliant from an engineering perspective”, says Alcherio Martinoli, a roboticist at the Swiss Federal Institute of Technology in Lausanne. “It just decouples a complex reverse-engineering problem into two pieces of information which have to work together,” says Martinoli, who was not involved in the work.

The remarkable ability of ants1 to work industriously without a leader was already noted by King Solomon:

לֵֽךְ־אֶל־נְמָלָ֥ה עָצֵ֑ל רְאֵ֖ה דְרָכֶ֣יהָ וַחֲכָֽם׃ אֲשֶׁ֖ר אֵֽין־לָ֥הּ קָצִ֗ין שֹׁטֵ֥ר וּמֹשֵֽׁל׃ תָּכִ֣ין בַּקַּ֣יִץ לַחְמָ֑הּ אָגְרָ֥ה בַ֝קָּצִ֗יר מַאֲכָלָֽהּ׃2

The Talmud records that R. Shimon b. Halafta, who apparently had a scientific mind and took nothing for granted, performed an experiment to determine whether ants really have no king, on account of which he was given the (possibly pejorative) appellation עסקן בדברים, “since he did not wish to rely upon the authority of Solomon” (Rashi):

הוה אמרו עליו על רבי שמעון בן חלפתא שעסקן בדברים היה … מאי עסקן בדברים [רש”י: למה נקרא שמו עסקן תחלה] א”ר משרשיא דכתיב לך אל נמלה עצל ראה דרכיה וחכם אשר אין לה קצין שוטר ומושל תכין בקיץ לחמה אמר איזיל איחזי אי ודאי הוא דלית להו מלכא [דכתיב אשר אין לה קצין והיינו עסקן דלא בעי למיסמך אהימנותיה דשלמה] [הנמלים אוהבים צל ושונאים חום] אזל בתקופת תמוז פרסיה לגלימיה אקינא דשומשמני [נמלים] נפק אתא חד מינייהו [מתוך חורי הקן ובא לחוץ וראה הצל] אתנח ביה סימנא [להכירו לידע מה יעשו לו חבריו כשימצא שקרן] על אמר להו נפל טולא נפקו ואתו דלייה לגלימיה נפל שמשא נפלו עליה וקטליה אמר שמע מינה לית להו מלכא דאי אית להו הרמנא דמלכא לא ליבעו3

The Talmud, however, proceeds to deny the conclusiveness of this experiment, concluding that we have no choice but to rely upon the authority of Solomon, “whose statement was inspired by רוח הקודש” (Rashi):

א”ל רב אחא בריה דרבא לרב אשי ודלמא מלכא הוה בהדייהו א”נ הרמנא דמלכא הוו נקיטי [שכל המהתל בחברו יהרגוהו] אי נמי בין מלכא למלכא הוה [מת המלך ועדיין לא קם אחר ודרך ארץ לעשות איש הישר בעיניו כדכתיב בימים ההם וגו’ איש הישר בעיניו יעשה] דכתיב בימים ההם אין מלך בישראל איש הישר בעיניו יעשה אלא סמוך אהימנותא דשלמה [נאמן הוא על כך וברוח הקדש אמרה]

The commentators note that the idea that it is legitimate to submit claims of the Sages to independent empirical verification apparently contradicts a different Talmudic account:

כי הא דיתיב רבי יוחנן וקא דריש עתיד הקב”ה להביא אבנים טובות ומרגליות שהם שלשים על שלשים וחוקק בהן עשר על עשרים ומעמידן בשערי ירושלים לגלג עליו אותו תלמיד השתא כביעתא דציצלא לא משכחינן כולי האי משכחינן לימים הפליגה ספינתו בים חזא מלאכי השרת דיתבי וקא מינסרי אבנים טובות ומרגליות שהם ל’ על ל’ וחקוק בהן עשר ברום עשרים אמר להו הני למאן אמרו ליה שעתיד הקב”ה להעמידן בשערי ירושלים אתא לקמיה דרבי יוחנן אמר ליה דרוש רבי לך נאה לדרוש כאשר אמרת כן ראיתי אמר לו ריקא אלמלא (לא) ראית לא האמנת מלגלג על דברי חכמים אתה נתן עיניו בו ונעשה גל של עצמות4

The Tosafos explain that R. Shimon was not challenging King Solomon’s assertion, but merely attempting to ascertain the source of his knowledge:

אע”ג דכה”ג חשיב בהמוכר הספינה (ב”ב עה.) מלגלג גבי ההוא תלמיד דאמר ליה רבי יוחנן אם לא ראית לא האמנת מלגלג על דברי חכמים אתה שאני הכא שבא לברר הדבר ולהודיע איך ידע שלמה:5

Maharsha considers this דחוק, since Solomon certainly derived his knowledge from רוח הקודש. He explains instead that there is an epistemological imperative to independently derive even those truths that are established by tradition:

כתבו התוספות … והוא דחוק דשלמה ודאי ברוח הקודש ידע ואמרה אבל הנראה דאין זה מלגלג כמו שכתב בעל חובות הלבבות בדברים התלוים באמונה אף שידע אותם בקבלה צריך כל אדם לחקור אותם במופת והוא שקראו עסקן בדברים בעלמא כגון זו עסק וראה ר”ש בן חלפתא לחקור אחריהם לידע אותם במופת וק”ל:6

Others explain that R. Shimon was simply fulfilling King Solomon’s instructions to “go to the ant [and] see …” – and so he said “I will go and I will see …!”

שמעתי מהרר”י סג”ל ז”ל דשלמה אמר לך אל כו’ ראה כו’ אמר איזיל אחזי כו’ וק”ל:7

Rav Yosef Ḥayyim of Baghdad explains that while one is indeed generally required to accept assertions of King Solomon as fact, משלי (where our discussion of the ant is found) is an exceptional case, since the work, as per its self-description, comprises parables, which may have been composed with poetic license and not be true in a literal sense:

הנה בודאי יש לו לסמוך על דברי שלמה המלך ע”ה ורק מפני שפסוק זה נאמר במשלי שהוא דברי משל וככתוב בתחלת הספר משלי שלמה בן דוד, לכן נסתפק אם הוא כן בודאי שאין להם מלך, או דלמא נקיט כן בדברי משל וצחות לשון ואין הדברים כפשוטן, לכך אמר איזיל אחזי אי ודאי הוא דלית להו מלכא:8

  1. Termites are not actually ants, but there are many similarities between them. []
  2. משלי ו:ו-ח. ועיין פה []
  3. חולין נז:‏ []
  4. בבא בתרא עה.‏ []
  5. תוספות חולין שם ד”ה איזיל ואחזי היכי איתא מילתא []
  6. מהרש”א חולין שם חידושי אגדות ד”ה מאי עסקן בדברים []
  7. מהר”ם שי”ף חולין בתוספות שם []
  8. בן יהוידע חולין שם ד”ה אמר איזיל אחזי []

Women Behind the Wheel

The internet is currently aflame with debate over (just fired) Google engineer James Damore’s “anti-diversity” memo or manifesto (Gizmodo; Motherboard; Medium), containing the incendiary suggestion that:

Differences in distributions of traits between men and women may in part explain why we don’t have 50% representation of women in tech and leadership.

and the corollary that:

Discrimination to reach equal representation is unfair, divisive, and bad for business.1

This post considers one of the most curious instances of an apparent Talmudic assumption of a “difference in distribution” of a particular characteristic between men and women (although as we shall see, the nature of the characteristic in question is unclear), and its concomitant halachic significance.

The Talmud rules:

ת”ר השוכר את החמור לרכוב עליה איש לא תרכב עליה אשה אשה רוכב עליה איש ואשה בין גדולה ובין קטנה אפילו מעוברת ואפילו מניקה2
בבא מציעא עט:‏

The clear implication is that female riders place more strain, or cause more wear, on donkeys than do male ones, although the Talmud, and (so far as I am aware) the medieval commentaries, are silent as to exactly why this is the case. The Sema gives the startling interpretation that “[a woman] is heavier [than a man]”:

לא ירכיב עליה אשה. שהיא כבידה:3

This explanation is problematic, on two counts. Firstly:

“Males weigh about 15% more than females, on average. For those older than 20 years of age, males in the US have an average weight of 86.1 kg (190 lbs), whereas females have an average weight of 74 kg (163 lbs).”

It is possible (actually, very probable) that this is at least somewhat culturally dependent. Even if we assume that there’s a natural, inherent disparity in average weight between men and women, this can surely be affected by factors in which cultural variations between genders exist, and vary across societies and eras, such as diet, physical activity and aesthetic preferences about body image. Perhaps, then, in the Talmudic era women actually were heavier, on average, than men, even though this is not the case today. [We know, after all, that women of that period actually preferred to look “fleshy”4, in stark contrast to the modern Thin Ideal.] If this is the case, however, we would expect poskim to note that these halachos do not apply, or actually apply in reverse, in contemporary times.

Secondly, even if we do assume that women of the Talmudic era were indeed heavier than men, this was presumably only true of the average, but surely there were many men who weighed more than many women. Since one who rents his donkey for use by a man is (presumably – see below) willing to allow any man to ride it, he should also be presumed to allow those women who weigh less than men to ride it. To this we can counter that the halachah establishes general rules, even if they sometimes result in illogical consequences. Since in general women are (or were) heavier than men, we presume that one who rents his donkey for use by a man means to limit it to such use, even though some of the allowed riders will be heavier than some of the disallowed ones.

In any event, this second objection was raised by the eighteenth century Italian Rav David Pardo:

ובטעמא דלא ירכיב אשה לא פירש”י ולא מידי ובחו”מ סימן הנזכר פירש הסמ”ע לפי שהיא כבידה ואני שמעתי ולא אבין מאי פסקא דאטו לא משכחת איש גדול ושמן ואשה קטנה וכחושה ואפילו הכי סתמא תנא לא ירכיב אשה אפילו קטנה

ונראה דהאי טעמא משום דסתם אשה אינה בקיאה ברכיבה ואינה יודעת וגם אינה בת דעת כל כך להשגיח להנהיגו יפה בדרך כבושה ושמא יתקלקל בהליכתו בדרך עקלתון או כיוצא בזה וכן נראה שהבין ריב”א הביאו המרדכי בפרק הנזכר וז”ל

פסק ריב”א דכל שכן אם הרכיב עליו גוי דגרע מאשה כי אינו חס על ממונו של ישראל ע”כ

מזה נראה דאשה נמי מהאי טעמא הוא לפי שאין לה דעת להקפיד ולהשגיח לחוס עליו שלא יוזק ומשום הכי אתי שפיר דאשה מרכיב עליה איש ואפילו גדול וכבד דהא מדקאמר אשה סתם היה יכול להרכיב אפילו גדולה וכבדה ביותר אם כן כשמרכיב איש על כל פנים עלויי קא מעלי ליה מטעמא דאמרן …5

In other words, R. Pardo understands that women are poorer drivers than men: they are less proficient, and have inferior judgment.

A major ramification of the varying interpretations of the Sema and R. Pardo is the halachah’s relevance to automobile rentals. According to the Sema, it would presumably not apply, since the relatively minor weight difference between males and females is insignificant in the context of automobiles. According to R. Pardo, however, the situation is less clear. Conventional wisdom stereotypes women as bad drivers, but the evidence is actually somewhat ambiguous. On the one hand, the statistics show that in general, male drivers cause more accidents than do female drivers. But on the other hand, this may be wholly or partially due to the fact that the former drive far more than the latter. Furthermore, at least one famous study found that women are actually overrepresented in certain types of crashes (although exactly why this is so is not entirely clear, and may even be caused by “common stereotypical expectations” that the drivers have of each other):

While men and women often disagree about which gender has better driving skills, a new study by the University of Michigan may shed some light on the debate.

Using data from a nationally representative sample of police-reported crashes from 1988 to 2007, Michael Sivak and Brandon Schoettle of the U-M Transportation Research Institute studied the gender effects in six different crash scenarios (based on crash angles, direction of approach and speed). These two-vehicle crash scenarios included various maneuvers in which one vehicle turned in front of the other, one vehicle side-swiped the other or both vehicles collided head-on. …

Sivak and Schoettle compared the actual frequencies of different combinations of involved male and female drivers in the six crash scenarios with the expected frequencies if there were no gender differences. The expected frequencies were based on annual distance driven for personal travel by male and female drivers. Because men drive about 60 percent of those annual miles and women drive 40 percent, men are expected to be involved in a higher percentage of crashes for each scenario, road conditions and driving skills being equal.

But the researchers found that crashes involving two female drivers were overrepresented in five of the six crash scenarios, including two by at least 50 percent more and two others by more than 25 percent greater than what was expected. On the other hand, crashes involving two male drivers were underrepresented in four of the six scenarios, including two by more than 20 percent and another by just less than 20 percent. In crash scenarios involving both male and female drivers, actual frequencies tended to be close to the expected frequencies.

“The results indicate that in certain crash scenarios, male-to-male crashes tend to be underrepresented and female-to-female crashes tend to be overrepresented,” Sivak said.

“This pattern of results could be due to either differential gender exposure to the different scenarios, differential gender capabilities to handle specific scenarios or differential expectations of actions by other drivers based on their gender.”

In all, success in handling on-road conflicts depends not only on psychomotor ability but also on the outcome of complex social interactions between traffic participants. In turn, these interactions are influenced by expectations based on prior experience “and a set of common stereotypical expectations that drivers have concerning the behavior of male and female drivers.”

A mini-haburah that I delivered a couple of years ago (partially) on this topic is available at the Internet Archive. A follow-up post will, בג”ה, deal with the propriety of women driving in general, beyond the context of rented conveyances.

  1. This controversy is reminiscent of the one of a decade ago over Larry Summer’s hypothesis (inter alia) that “different availability of aptitude at the high end” may explain “the very substantial disparities … with respect to the presence of women in high-end scientific professions.” []
  2. בבא מציעא עט:‏ []
  3. סמ”ע ריש סימן ש”ח []
  4. שבת ריש נז:‏ []
  5. חסדי דוד (חלק שני) שם פרק ד’ (עמוד לג.) ד”ה השוכר את החמור להרכיב איש []

The Army Corps Of Engineers Plays God

Yesterday’s edition of NPR’s The Takeaway discussed the “devil’s choice” facing the Army Corps of Engineers with regard to opening the Morganza Spillway on the flooding Mississippi River; as Kavanaugh Breazeale, Public Affairs Officer for the Vicksburg District Army Corps of Engineers explains the dilemma:

It is unfortunate that you flood farmlands and timberlands but it benefits the residential areas, that may be affected by high water.

Or as George Sills puts it:

It’s a very tough decision to put water in people’s homes and lives

Senior Meteorologist Jonathan Erdman, Sr. explains at Weather.com:

Morganza Spillway: Flooding Some to Save Many

One additional spillway is designed to be opened only during extreme Mississippi River flood events that threaten to compromise the Old River Control Structure (which, would then allow the swollen Mississippi River to jump to the Atchafalaya River channel).

Upstream from Baton Rouge near the town of Morganza lies the Morganza Spillway. …

Built in 1954, the control structure and floodway have only been put into action once, during the 1973 flood.

Unlike the Bonnet Carre, the decision to open the Morganza Spillway is complicated by its impact on populated areas of the Atchafalaya River Basin. According to a U.S. Army Corps inundation map, various towns and rural areas normally dry would be inundated by at least 5 feet of water upon opening the Morganza Spillway!

This is clearly one of the toughest decisions made by the U.S. Army Corps of Engineers. If you don’t open the spillway…

  • You will still have river flooding along the Atchafalaya River, just not nearly as expansive.
  • You risk putting undue pressure on the Old River Control Structure, which may get overwhelmed.
  • This would cause the catastrophic “jumping” of the Mississippi River to the Atchafalaya River Basin, as described earlier.
  • Pressure on levees in Baton Rouge and New Orleans would remain high, if the Old River Control Structure does not fail.

However, if you do open the spillway…

  • There will be widespread inundation of the Atchafalaya Basin.
  • Areas outside of levee protection would be under at least 5 feet of water.
  • Pressure on levees in Baton Rouge and New Orleans would be lower.

Sadly, the choice here may be to open the Morganza floodway, flooding the Atchafalaya Basin to avoid the potential of a more catastrophic re-routing of the Mississippi River.

Louisiana Governor Bobby Jindal put the risk very simply:

“If you were flooded in ’73, you’re likely to get wet again. If you nearly got wet in 1973, you’ll probably get wet this time.”

This is strikingly close to the classic scenario discussed in the Yerushalmi:

ראה אמת המים שוטפת ובאה לתוך שדהו רשאי לפנותן למקום אחר משנכנסו אין רשאי לפנותן למקום אחר1

with the potentially significant difference that the spillway is under the control of the Army Corps of Engineers, a public, ostensibly neutral third party, rather than in the hands of one of the interested parties. For a discussion of the literature surrounding this Yerushalmi, see our Seforim blog essay: Hunted Bears, Cantonists and Nazi Victims.

Update: the Corps has apparently decided to open the Spillway.

  1. בבא קמא פרק ג’ הלכה א []