Hostname: page-component-8448b6f56d-xtgtn Total loading time: 0 Render date: 2024-04-18T01:03:47.496Z Has data issue: false hasContentIssue false

The Origin and Date of the ‘Sator’ Word-Square

Published online by Cambridge University Press:  25 March 2011

Donald Atkinson
Professor of Ancient History in the University of Manchester


The recent extensive article on this subject by the distinguished French Scholar Dr. Jérome Carcopino provides an opportunity to return to a subject which I have already discussed in a paper published shortly before the War. Since its publication fresh evidence has come to light which may be thought to bear upon the subject, and various explanations of the origin of the square have been put forward which seem to invite comment. The exhaustive article of Fr. de Jerphanion, which is largely summarized by Carcopino, makes necessary only the briefest sketch of the development of our knowledge of the square. In its later form (beginning with Sator, fig. 2) it can be traced in a more or less complete form from the sixth century to modern times over an area extending from France to Ethiopia, Nubia, and even to South America, where its prophylactic virtues were accepted in the nineteenth century. The discovery of its earlier form (beginning with Rotas, fig. 1) incised on Roman wall-piaster at Cirencester in 1868 caused little interest, and it remained almost exclusively a matter for the medievalist for many years. In 1926 it was observed by Grosser that twenty-one of the letters of the square were made up of the word ‘Paternoster’ twice over (but necessarily arranged as a cross since the N, the middle letter, appears only once) and that the remaining four letters are two A's and two O's (fig. 4).

Research Article
Copyright © Cambridge University Press 1951

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)


page 1 note 1 Le Christianisme secret du “carre magique”’, Museum Helveticum, v (1948), 167 f.Google Scholar

page 1 note 2 The Sator Formula and the Beginnings of Christianity’, Bulletin of the John Rylands Library, xxii (1938), 419.Google Scholar

page 1 note 3 La Formule magique Sator arepo ou rotas opera: vieilles theories et faits nouveaux’, Recherches de science religieuse, xxv (1935), 188 f.Google Scholar

page 1 note 4 Op. cit., 217.

page 2 note 1 Cf. Arch. Journal, lvi (1899), 320Google Scholar; Ephaneris Epigraphica ix, 1001. It has become customary to date this Cirencester example to the third century A.D. (cf. Carcopino, op. cit., 25, following Rostovtzeff (Doura Prelim. Rep., v (1934), 159 f.Google Scholar)), but there is in fact no reason why it should not belong to the second or the fourth. It may be remarked here that Carcopino's suggestion of a military origin (op. cit., 56 n. 170) is fallacious. There is no evidence that Cirencester was ever a legionary station(the reference to Haverfield, Roman Occupation of Britain, 104, is a confusion with Wroxeter or Chester), nor, except perhaps for a short time under Claudius or Nero, anything but a civil town.

page 2 note 2 Ein neuer Vorschlag zur Deutung der Satorformel’ in Archiv. für. Religionswissenschaft, xxix (1926), 165 f.Google Scholar

page 2 note 3 He was not aware of the existence of the Cirencester example.

page 2 note 4 The relevant text is Barnabas, Ep. ix, 7–8: ‘And because the Cross in the Tau signifies (the grace) of God… etc’

page 2 note 5 By Della, M. Corte (its discoverer) in Rendiconti dell'Accad. Pontif. di Arch. Rom., xii (1936), 397 f.Google Scholar

page 2 note 6 Not. Scav. (1929), 449.

page 2 note 7 My previous article was written in 1936–7 before the publication of the second example, though it was not published till 1938.

page 3 note 1 In Contesrendus Acad. Inscript. (1937), 87 f.: and in much the same terms in Rendiconti dell'Accad. Pontif. di Arch. Rom., xix (1936), 401 f. (published in 1937).Google Scholar

page 3 note 2 Ezek. i. ix. x.

page 3 note 3 It will be observed that if this is so the Christian origin of the formula is still maintained, and I have not found any evidence that Cumont receded from this opinion.

page 3 note 4 ‘Une nouvelle hypotese sur l'origine du Carré magique “Rotas Opera”’, La Voix des Monuments, Nouvelle Série, 90 f. = Recherches de science religieuse (1937).

page 3 note 5 Corpus Inscriptionum Iudaicarum, i, Index s.v. γραματες but this interpretation of their functions is at least doubtful; may they not rather be officials of the Synagogues (? Scribes).

page 3 note 6 i. 4–17.

page 4 note 1 In the Vulgate text.

page 4 note 2 I.e. the Jewish letter corresponding to T (Greek Tau) and the last letter of the Hebrew Alphabet. Its Jewish symbolism is chiefly connected with the Torah—the Law.

page 4 note 3 Ezek.ix. 1–6.

page 4 note 4 Cumont pointed out (Rendic. Accad. Pontif. di Arch. Rom., xiii (1937), 7 f.Google Scholar) that the Septuagint word translated ‘scatter’ (Vulg. effunde) is διασκρπισον =sow; but it is doubtful whether this interpretation of the word is justified, while his statement that in Ezek. i. 16 (in the Vulgate) the wheels are twice connected with the word opera, fails to notice that here it is neuter plural (from Opus), not ablative singular (from Opera), which alone makes sense in the rebus, and that in any case this would imply a Latin translation of Ezekiel at a very early date.

page 4 note 5 Ezek. x. 1–2.

page 4 note 6 On these points cf. Ezechiel, ed. J. Herrmann, 1924; Hesekiel, ed. Bertholet, 1936; La Grands Prophetes, ed. Dennefeld, 1946; Ezechiel, ed. F. Spadafora, 1938; Ezechiel, ed. P. Huvray, 1949; Ezechiel, J. Ziegler, 1948; references which I owe to my colleague Professor H.H.Rowley.

page 4 note 7 Op. cit., 9. Cf. Carcopino, op. cit., 40.

page 4 note 8 Rendic. Accad. Pontif. di Arch. Rom., xii (1936), 409Google Scholar: ‘Cosi la parola Arepo messo in apposizione a Sator non e se non un nome proprio inventato per le esigenzc stesse del quadrato, leggendo a rovescio la parola opera. E puro caso se in essa si trova una radice celtica.’

page 5 note 1 See p. 4, note 4 above.

page 5 note 2 On this cf. Carcopino, op. cit., 41.

page 6 note 1 In C.R.A.I. (1937), 87 f.

page 6 note 2 On this, see below, p. 8.

page 6 note 3 For this is the date of the earliest known example of the Sator arepo arrangement (Fig. 2), which Jerphanion takes to be the Christian form.

page 6 note 4 Besides this Jewish hypothesis of the origin of the square, two other interpretations have been suggested. Omodeo (Critica, xxviii (1940), 45Google Scholar), sees the square as Mithraic; Mithras is the Sator and the wheels are those of the Solar chariot: Sundwall (Acta Acadtmiae Aboensis, Humaniora, xv. 5 (1945), 1617Google Scholar), regards it as Orphic, Eleusinian; Sator is Triptolemus and the wheels represent the plough which is one of his attributes. On these see Carcopino, op. cit., 38–9, to whom these references are due. There is clearly a serious chronological difficulty about the first, and both suffer from a great lack of inherent probability. It does not appear that either has met with a favourable reception.

page 8 note 1 For the latter's arguments, sec above, p. 3.

page 8 note 2 Acts xxviii. 13 f.

page 8 note 3 Tertullian ApoL, xl. 8. See below, p. 15.

page 8 note 4 On this, see below, p. 16.

page 8 note 5 Not. Scav. (1937), 177. It b worth remarking that a recent paper by Dr. O. Onorato (Accad. Linen Rendic. Class. Sc. Morali, etc. Série viii, iv, 644 f.) argues convincingly for A.D. 62 as the date of the earthquake.

page 8 note 6 In the discussion following Jerphanion's rejection of the Christian origin in the paper cited above, p. 3, n. 1.

page 9 note 1 ‘Esplorazioni di Pompei immediatamente successive alia catastrofe dell' anno 79’, in In Memoria Vasile Parvan (1934), 96–109. The paper was reprinted with some additional material in Historia (Aprile-Giugno 1934), No. 2, Anno viii. 354.

page 9 note 2 Reg. vii; Ins. ii. 20: Delia Corte, Case ed Abitanti, no. 189.

page 9 note 3 Fiorelli, Dtscrizionc di Pompei, 190: C.I.L., iv. 2311.

page 9 note 4 Nos. 6825, 6820.

page 9 note 5 Cf. Historia, I.c., fig. 4.

page 9 note 6 C.I.L., iv.6819.

page 10 note 1 The letters restored are not broken away, but were never written.

page 10 note 2 C.I.L., iv. 6826.

page 10 note 3 C.I.L., iv. 6821.

page 10 note 4 Reg. I, Ins. I. The graffiti are published by Delia Corte in Not. Scav. (1929), 448 f.

page 10 note 5 C.I.L., iv. 4976.

page 11 note 1 Genesis xiii. 10–19, 29.

page 11 note 2 Carcopino argues that its form (Latin in Greek letters) makes it necessary to date it to the end of the second century at the earliest, and cites some instances of the practice not earlier than the third century. But since there are, as he admits, other instances from Pompeii which he does not claim to be of post-eruption date, his argument is inconclusive; nor is it of great significance even if it were accepted.

page 11 note 3 The greater part, but not the whole, of the open courtyard has been cleared.

page 11 note 4 Not. Scan. (1939), 165 f.

page 11 note 5 Sogliano, Pitture murali (1879), n. 604. Delia Corte has recently shown reason to suppose that a more correct name for the building would be ‘campus’: Accad. Naz Linei Rendic.Classc Sc. Mor.ecc., Série viii, ii (1947), 555f.

page 12 note 1 Twenty-four alphabets, complete or partial, and other graffiti indicate this: Not. Scav. (1939), 241 f., nos. 2, 45, 52, 55–6, 84–5, 94, 99, 129, 142, etc.

page 12 note 2 During a recent visit to Pompeii I had the advantage of a discussion on this point with Dr Delia Corte, whose practical acquaintance with Pompeian excavations covers nearly half a century.

page 12 note 3 Cf. Pliny, Epp., vi. 16.

page 13 note 1 Op. cit., 20 n. 31, 52 nn. 144–7, where he cites De resurr, carn., viii; De. cor. mil., iii; Ad ux., ii, 5; Adv. Marc, iii. 22.

page 13 note 2 v. 1. 6: Galli semi-iugerum quoque arepennem vocant.

page 13 note 3 Isidore of Seville (c. A.D. 600), (xv. 15; ed. Lind p. 486), remarks that in his own province of Baetica the same word is used (in the form arapennis) with the same meaning. The word therefore might well have reached both Gaul and south-west Spain from some Italian dialect and may not be Celtic at all.

page 13 note 4 N.H., xviii. 172: Non pridem inventum, in Raetia Galliae est duas addere tali (vomeri) rotulas quod genus Raeti vocant plaumorati. Raetia Galliae has been interpreted as being the Vallis Poenina, the Rhone valley above Geneva, the four communities of which had been taken away from Raetia and added to Gaul by Claudius.

page 13 note 5 On this cf. Carcopino, op. cit., 28 f.

page 14 note 1 Op. cit., 22 f., cf. 58.

page 14 note 2 Cardon; Rerum Validate, cited by Jerphanion, op. cit., 213.

page 14 note 3 Op. cit., 200.

page 14 note 4 Sador, Alador, Danet, Adera, Rodas, i.e. the later form of the square.

page 14 note 5 Irenaeus, Adv. Haer., ii. 24, 4. Presumably the five summitates are the five salient points of the Cross, one for the feet, two for the hands of Christ, one at the junction of the beams, and one to fix the writing at the top.

page 14 note 6 Bulletin di la Soc. nat. des Antiquairus de France (November 1934), 7.

page 15 note 1 P. 3f.

page 15 note 2 xl. 8.

page 15 note 3 Pliny, N.H., ii. 139–40, records that Vobinii was completely destroyed by lightning in the reign of King Porsena (sixth century B.C.).

page 15 note 4 How many people in England or France could give off-hand the dates of the Lisbon earthquake or the destruction of Messina?

page 15 note 5 Cf. above, p. 8.

page 15 note 6 On this see Maiuri L' Ultima Fase Edilizia di Pompeii, e.g. Tav. lviii–lxi.

page 15 note 7 Pliny, N.H., xxxi. 94.

page 15 note 8 E.g. C.I.L., iv. 2152, a greeting to the Colony of Puteoli presumably written by a Puteolan.

page 16 note 1 Les Ruines de Pompei, 2me partie (Paris 1824), 84 f.Google Scholar: ‘Sur un panneau de stuc blanc, une especc de croix en bas-relief.’ Mazois' illustration is repeated by Corte, Delia, ‘I Cristiani a Pompei’, in Rendic. d. R. Accad. di Arch. Lett, e Belle Arti delta Soc. R. di Napoli, xix (1938), 25 (of the off-print).Google Scholar

page 16 note 2 C.I.L., iv, 679. Cf. Delia Corte, op. cit., 6 f. The original soon faded and the variation between the two copies made ((a). ristiani, (b) Christianos) has naturally been emphasised by the sceptical.

page 16 note 3 Rendic. dell' Accad.pontif. di Arch. Rom., xv. 193 f. (with plans and photographs).

page 17 note 1 Op. cit., 38.

page 17 note 2 Matt. vi. 6.

page 17 note 3 Op. cit., 38.

page 17 note 4 Le Crux interpretum di Ercolano’, Rendic. dell' Accad. pontif. di Arch. Rom., xxi (1945–6), 15.Google Scholar

page 18 note 1 Byzmtion ii (1925), 337 f.