We have seen that ancient Judaism associated the act of writing as well as the objects on which letters and symbols appear with power and protection. Amulets and gems inscribed with angelic names, unpronounceable formulas, voces magicae, and portions of Scripture provided ordinary materials with a magical essence. Knowledge is found in writing, and strange writing, especially to an illiterate person, holds even deeper and stranger knowledge.
I propose that (certain) coins were perceived as magical by Jews in Late Antiquity as well. I come to this conclusion for several reasons. First, Late Roman and Byzantine coins were inscribed with both written inscriptions and images that were often enigmatic to the people encountering them. As we have seen, Late Roman-Byzantine coins were inscribed with legends and fieldmarks, often abbreviated to a couple of symbols or markers. The legends were generally written in Latin, a language that the average Jew in Palestine could not read nor speak. According to William Schniedewind, writing in ancient Judaism had a numinous power. He states “Writing was not mundane; rather, writing was used to communicate with the divine realm by ritual actions or formulaic recitations in order to affect the course of present or future events.” Writing was a gift from God and had the supernatural power to curse or to bless. Besides the inscriptions, coins also had finely minted images on them, depicting emperors but also angels, animals, crosses, Victory, and other symbols. Often these emblematic signs were depicted deliberately for their ritual symbolism, such as the Christogram, which appears on coins from Constantine the Great onwards as well as on amulets and papyri with magical spells. All these elements gave coins a mysterious allure, and it would not be unreasonable to suppose that people perceived coins and amulets as enigmatic devices, classifying them under the materia magica. In fact, coins with holes drilled into them are found around the ancient Mediterranean world, and are interpreted as repurposed coins to be worn on the body as jewelry or amulets. Literary sources from Late Antiquity, both Jewish and non-Jewish, confirm this practice. According to Henry Maguire, empresses, dependent rulers, and high court officials in Late Antiquity and the early Medieval period often wore portraits of the reigning emperor woven into or sewn onto their garments. Like diplomatic gifts, the images of the emperor displayed upon the person were not only a sign of the emperor’s rule but a conduit of his protection. Sometimes, however, it is clear from excavations that the drilled coin was already old at the time when it was worn, suggesting that its apotropaic value could have been increased by its age. John Chrysostom, for example, living in the 4th century CE, scolded the superstitious who wore the coins of Alexander the Great — a ruler who lived 700 years earlier — as periapta or amulets. It seems unlikely that Jews in Palestine sought imperial protection by wearing coins with images of Roman or Byzantine emperors on their body, although a coin with a hole drilled into it was found in the synagogue at Horvat Kur (a 5th century 40 nummi depicting Justinian I, minted in Antioch). Instead, I propose that coins found in walls, behind benches, under the floor, and in the foundations of ancient synagogues might have been placed there as apotropaic devices.
Magic in the ancient world was often connected with the use of metals. Besides being inscribed with symbols and powerful images, coins were made of metal as well. As we have seen, writing magical texts on sheets of gold, silver, bronze, lead, or tin (lamellae) was common in the Greco-Roman world. In Sefer ha-Razim, a collection of magical texts, probably collected and written in the 4th century CE, multiple recipes call for writing Jewish magical inscriptions on a metal surface. For example, magical texts are supposed to be inscribed on bronze lamellae in recipes I/201, II/31, II/115-116, II/151-152; on tin lamellae in recipe I/144; on lead in recipe II/63-64; on iron in recipe II/111-112; on silver in recipes II/54-55, II/100-101, II/126-127, II/137-138, III/37; and on gold in recipes II/125-126, V/20, VI/29-31. These are not the only instances of metals being used for magical devices in ancient Judaism. The Babylonian Talmud contains the following recipe against fever:
“For a daily fever…he should sit at a crossroads and when he sees a big ant carrying something, let him take it and place it within a copper tube, close it with lead, seal it with sixty seals, shake it, lift it up, and say to it [the ant]: “Your burden upon me and my burden upon you.” (bt. Shabbat 66b)
Some Aramaic incantation bowls also refer to metals and even coins to protect oneself against evil. Thus, we can conclude that coins in Late Antiquity could transcend the market place and appear in numerous other contexts, including magical ones. To be clear, this idea is not new. In 1976, Yaakov Meshorer proposed that the 139 coins from the 1st century, found hidden in an oil lamp in the walls of a house at En-Gedi were placed there as a protective measure. However, based on all the above evidence, I believe that the coins placed under the floors or in the walls of ancient synagogues as well can be interpreted as apotropaic devices. But does this make these coins “foundation deposits”, as many scholars have concluded?
 In this they were not alone but part of a broader pan-Mediterranean phenomenon in which writing or meaning-making of the world were seen as magical (Frankfurter 1994). As for Jewish magic, the majority of discovered magical recipe texts ask for the inscribing of a particular text on various surfaces such as on parchment, potsherds, or eggs (Saar 2017, p. 90). A well-known example of the power of writing can be found in Numbers 5:16-30, which portrays the so-called sotah-ritual: A priest brings a woman accused of adultery before YHWH and makes a potion in which the main ingredient is writing: “Then the priest shall put these curses in writing, and wash them off into the water of bitterness. He shall make the woman drink the water of bitterness that brings the curse, and the water that brings the curse shall enter her and cause bitter pain… and afterward he shall make the woman drink the water.” The writing might have been on a piece of papyrus dissolved in the water, or an ostracon with letters which that were washed off in the water: in any case, it is the writing that gave the water power (if the ritual was actually practiced or if it was only ideological is beside the point; the story is still an indication of the power of the written word. See Rosen-Zvi 2012).
 A potsherd found in the fill of a room north of the Horvat Rimmon synagogue confirms this. The sherd, dated to the 5th or 6th century, contains eight lines of text and one line of “magic characters.” The sherd does not seem to be an accidental fragment, but the potter had deliberately cut deep incisions in the pot before firing. This way, the jar could be broken along these lines after the clay had hardened. The writing as well was made before the clay was baked. The text invokes angels and has parallels among the Cairo Genizah scrolls (Naveh and Shaked 1985, p. 87; Kloner 1989, p.47; Swartz 2018, pp. 105–109).
 This does not per se mean that the average ancient Jew could not read or interpret coins in general, only that one had to be informed on the symbolic abbreviations and language before its knowledge could be revealed.
 Schniedewind 2004, pp. 24–34. The word numinous was coined by Rudolph Otto in 1917 in his book Das Heilige, as a term meaning “arousing spiritual or religious emotion; mysterious or awe-inspiring.” For more on writing as a magical act in ancient Judaism, see Stern 2018; 2021.
 Schniedewind 2004, p. 24.
 Maguire 1997, pp. 1038–1039, de Bruyn 2017, pp. 56–66. Another powerful image in Late Antique magic was the “Holy Rider”: a victory motif found on amulets associated with King Solomon and his power to defeat and control demons (Bonner 1950, pp. 208–221 and Plates 295–326; Fulghum 2001, p. 142; Russell 1995, pp. 40–41; Morrisson 2014). Could this motif also have been recognized in the Fallen Horsemen coins which look very much the same?
 Bendall 1995; Morrisson 2012; Syon 2015, p. 41; Winges 2017.
 Maguire 1997; Fulghum 2001; Rowan 2009.
 Maguire 1997, p. 1039.
 Maguire 1997, p. 1039, who mentions an example from the 13th century CE, when the Byzantine orator Hobolos gave a cloth embroidered with the image of the emperor as a gift to the city of Genoa, saying to the emperor “the form of the beloved [emperor], even in a picture, was a great remedy to those who love you. For even your image, if it is beside us, has many powers. It will be a firm means of defense against our adversaries, an averter of every plot, a strong bulwark for your city and ours…”
 For example, Fulghum 2001, pp. 143–144. However, we also know from rabbinic sources that some rabbis encouraged the use of “old” (“שֶׁנִּפְסַל”: invalidated, chipped, cut up) coins to be used as jewelry (m. Kel. 12:7).
 John Chrysostom, Ad illuminandos catechesis secunda, II, 52: “…What is one to say about those who use enchantments and amulets, and who tie bronze coins of Alexander of Macedon around their heads and feet?” It is clear that the image of Alexander the Great, a powerful ruler, would have been seen as very potent.
 According to Winges, the largest sample of perforated coins are Justiniac folles, further indicating that this coin was used as an amulet (Winges 2017, p. 13. See also Morrisson 2014).
 Sometimes coins were not drilled but simply set into a piece of jewelry, in which case we would not be able to recognize them as amulets (Bruhn 1993; Rowan 2009).
 See also Naveh and Shaked 1985; Kotansky 1994; Bohak 2008, pp. 149–153, Leiman and Leibner 2016.
 For the Hebrew text, see Margalioth 1966 and for English, see Morgan 1983. For an overview on the debate around its dating, see Bohak 2008, pp. 170–175.
 The Roman numeral refers to one of the seven “firmaments/heavens,” or books into which the Sepher is divided. The Latin numbers refer to the text lines.
 For example, bowl BM 91715 (https://www.britishmuseum.org/collection/object/W_1881-0714-3), stored at the British Museum (Ford 2002, pp. 32–36).
 Meshorer 1976, repeated in 2007.