No CrossRef data available.
Article contents
The Subscriptions to Mark's Gospel and History of Reception
Published online by Cambridge University Press: 15 November 2023
Abstract
This article surveys the subscriptions to the Gospel of Mark in 157 Greek manuscripts, noting their gradual development from being identical to the title in the earliest phase to becoming more and more elaborate and significant for the history of interpretation. Early on, as reflected in the title, the Second Gospel was associated with Mark, known to be Peter's disciple and interpreter. In the fourth or fifth century, an editor added the information that it was written (or spoken) by Mark in Latin in Rome as reflected in the Peshitta and later Byzantine manuscripts. At some point between the seventh and ninth centuries, an unknown editor added dates for each of the four Gospels from a source which has been attributed to Hippolytus’ Chronicle, and in the process made a cascading error which resulted in too early dates for Mark, Luke and John. In the archetype of Family 13, these traditions were combined which suggests that the archetype of Family 13 is no earlier than the eighth century. A main factor behind this gradual growth of the subscriptions is authentication and authorisation – in the case of the Second Gospel, the association with Mark and Peter legitimates its claim of apostolicity and orthodoxy. Moreover, the situating of each Gospel in time and space through the subscriptions not only satisfies human curiosity but contributes to the construction of an ancient Christian ‘landscape of memory’, reflecting the collective memory of the early Christians, thus shaping and enhancing their identity.
Keywords
- Type
- Articles
- Information
- Copyright
- Copyright © The Author(s), 2023. Published by Cambridge University Press
References
1 Novum Testamentum Graecum: Editio Critica Maior I: The Synoptic Gospels; 2: The Gospel according to Mark; 1: Text (ed. Holger Strutwolf et al.; Stuttgart: Deutsche Bibelgesellschaft, 2021). Our particular selection of Greek manuscripts was made not only to represent the textual tradition of Mark as a whole but also to evaluate the ECM edition of Mark. We acknowledge the fact that a full collation of the many Byzantine manuscripts would likely add a few more subscriptions (or alternative forms of existing ones) to our apparatus.
2 Hengel, Martin, ‘Entstehungzeit und Situation des Markusevangeliums’, Markus-Philologie: Historische, literargeschichtliche und stilistische Untersuchungen zum zweiten Evangelium (ed. Cancik, Hubert; WUNT 33; Tübingen: Mohr Siebeck, 1984) 1–45Google Scholar; Günther Zuntz, ‘Wann wurde das Evangelium Marci geschrieben?’, Markus-Philologie 47–71.
3 Hengel, ‘Entstehungzeit’, 9.
4 Zuntz, ‘Evangelium Marci’, 51–71.
5 Wasserman, Tommy and Thorp, Linnea, ‘The Tradition and Development of the Subscriptions to 1 Timothy’, Paratextual Features in Papyrology and Early Christian Manuscripts (ed. Porter, Stanley E., Stevens, Chris S., and Yoon, David I.; TENTS 16; Leiden: Brill, 2023) 172–201Google Scholar; and Tommy Wasserman and Conrad Thorup Elmelund, ‘Second Timothy: When and Where? Text and Traditions in the Subscriptions’, Paratextual Features, 202–26. Some terminology has been drawn from David G. Champagne, ‘Scribal Habits Within the Superscription and Subscription Traditions of Greek New Testament Manuscripts’, Ph.D. diss. (New Orleans Baptist Seminary, 2012), specifically the use of ‘modifier’ to designate various elements in the subscriptions, e.g., ɛὐαγγέλιον as ‘genre modifier’.
6 Hatch, Facsimiles and Descriptions of Minuscule Manuscripts of the New Testament (Cambridge: Harvard University Press, 1951) 33–4.
7 Hatch, Facsimiles, 33.
8 See Tommy Wasserman, ‘The Greek New Testament Manuscripts in Sweden with an Excursus on the Jerusalem Colophon’, SEÅ 75 (2010) 86–92. The colophon is first attested in Codex 565 and Codex Λ (039), both dated to the ninth century.
9 The ECM of Mark has 170 entries in the subscription apparatus. Our apparatus adds another twenty-five entries and removes three entries from ECM which should not have been included: 1689r occurs twice; 2542 is actually lacunose; 1582Cr records a scholion to Mark 16.19 which is not a subscription.
10 These forty-five manuscripts are: 05S, 011, 017, 019, 041S, 4, 18, 28, 61, 105, 117, 131, 153, 154, 205, 261, 273, 389, 513, 544, 565, 706, 752, 788, 872, 873, 892, 954, 979, 983, 1160, 1273, 1446, 1506, 1555, 1574S, 1582, 1675, 1689, 2174, 2193, 2206, 2542, 2886, 1387. In seven additional cases, the presented text is correct, but, in our opinion, the notation is wrong (this concerns 472, 695, 713, 828, 1009, 1253, 1654). The full transcriptions including proposed corrections to the ECM of Mark and other notes have been deposited publicly in the Dutch national centre of expertise and repository for research data (DANS): https://doi.org/10.17026/dans-zdf-b9sk.
11 Simon Gathercole, ‘The Titles of the Gospels in the Earliest New Testament Manuscripts’, ZNW 104 (2013) 33–76, 39.
12 Martin Hengel, Die Evangelienüberschriften (Heidelberg: Carl Winter, 1984) 47–51. Further, Hengel proposed that the genre designation ɛὐαγγέλιον derives from Mark 1.1 (p. 49).
13 Silke Petersen, ‘Die Evangelienüberschriften und die Entstehung des neutestamentlichen Kanons’, ZNW 97 (2006) 250–74, 274.
14 David Trobisch, The First Edition of the New Testament (New York: Oxford University Press, 2000) 38, 41.
15 Gathercole, ‘Titles’, 69. Gathercole has devoted a separate study to Papyrus 4, the flyleaf of which contains nothing but the title of Matthew's Gospel written as ‘ɛυαγγɛλιον κ̣ατ̣αμαθ’θαιον’, which he regards as the earliest extant manuscript title of Matthew. See Simon Gathercole, ‘The Earliest Manuscript Title of Matthew's Gospel (BnF Suppl. gr. 1120 ii 3 / 
4)’ NovT 54 (2012) 209–35. More recently, Brent Nongbri, God's Library: The Archaeology of the Earliest Christian Manuscripts (New Haven: Yale University Press, 2018) 247–68, has argued that this manuscript may be as late as the fourth century. See, however, the response by Tommy Wasserman who argues for a date around 200 ce in ‘Beyond Palaeography: Text, Paratext and Dating of Early Christian Papyri’, The Chester Beatty Biblical Papyri at Ninety: Literature, Papyrology, Ethics (ed. by Garrick V. Allen, Usama Gad, Kelsie Rodenbiker, Anthony Royle and Jill Unkel; Manuscripta Biblica 10; Berlin: Walter De Gruyter, 2023) 152–60.
16 Gathercole, ‘The Titles’, 69.
17 Simon Gathercole, ‘The Alleged Anonymity of the Canonical Gospels’, JTS 69 (2018) 447–76, esp. p. 454. Gathercole's main point in this study is that although the standardised titles were added in the second century (the consensus view) that certainly does not mean they were anonymous (or that the earlier Gospel manuscripts were unaccompanied by any indication of a name, e.g., an author's name could be written on the back of a roll or on a separate name tag).
18 Petersen, ‘Evangelienüberschriften’, 254; Gathercole, ‘The Titles’, 33–7. Although Trobisch is less clear on this particular matter, he does think that the long form is original to the ‘canonical edition’ (p. 126 n142). Moreover, he proposes that modern editions should include titles of ‘the literary unit of the Four-Gospel Book’, i.e., ɛυαγγɛλιοι δ’ (p. 103).
19 Vito Lorusso, ‘Locating Greek Manuscripts through Paratexts: Examples from the Library of Cardinal Bessarion and Other Manuscript Collections’, Tracing Manuscripts in Time and Space through Paratexts, (ed. Giovanni Ciotti and Hang Lin; Studies in Manuscript Cultures 7; Berlin, Boston: De Gruyter, 2016) 246.
20 Francesca Schironi, To Mega Biblion: Book-Ends, End-Titles, and Coronides in Papyri with Hexametric Poetry (The American Studies in Papyrology 48; Durham: The American Society of Papyrologists, 2010) 21–2, 164, 168 (P.Mil.Vogl. inv. 1225 and P.Lond.Lit. 5).
21 Cf. David C. Parker, Codex Bezae: An Early Christian Manuscript and Its Text (Cambridge: Cambridge University Press, 1992), 11, ‘[t]he use of explicit (i.e. explicitum) and incipit is standard in Latin books’.
22 For further examples of subscriptions outside our sample, see Hermann von Soden, Die Schriften des Neuen Testaments in ihrer ältesten erreichbaren Textgestalt hergestellt auf Grund ihrer Textgeschichte, 2 parts in 4 vols., 2nd unchanged ed. (Göttingen: Vandenhoeck & Ruprecht, 1911–3), I.1:297–9.
23 For the use of χρόνος as indication of a year, as attested in T. Jud. 7.10, Acts Paul 45 and elsewhere, see PGL s.v. χρόνος 3.
24 Hengel, ‘Enstehungszeit’, 8–9.
25 Hengel, ‘Enstehungszeit’, 9, ‘Die historische Wurzel dieser eigenartigen Tradition hängt gewiß auch mit der runden bzw. heiligen Zehn- oder Zwölfzahl zusammen . . .’
26 Zuntz, ‘Evangelium Marci’, 47. More recently, Nicholas H. Taylor, ‘Palestinian Christianity and the Caligula Crisis, Part 2, The Markan Eschatological Discourse’, JSNT 62 (1996) 13–41; and James G. Crossley, The Date of Mark's Gospel: Insight from the Law in Earliest Christianity (JSNTSup 266; London: T&T Clark, 2004) have argued along the same lines.
27 Zuntz, ‘Evangelium Marci’, 57–65.
28 Zuntz, ‘Evangelium Marci’, 66.
29 Zuntz, ‘Evangelium Marci’, 66–7.
30 Zuntz, ‘Evangelium Marci’, 65–71, esp. 70–1.
31 Franz Diekamp, Hippolytos von Theben: Texte und Untersuchungen (Münster: Aschendorff, 1898) 40.
32 Pinakes labels this untitled excerpt ‘Hippolytus Thebanus, Excerptum De Iacobo fratre Domini’, (https://pinakes.irht.cnrs.fr/notices/oeuvre/16467/), since the fragment follows after a section that commences, Ϊάκωβος δὲ γɛγονὼς πρῶτος ἐπίσκοπος on folio 323r. As evident from Georgi Parpulov's catalogue of catena manuscripts of the Greek New Testament, no other manuscript of the same catena-type (e.2.iii.α) attests to this text. See Georgi R. Parpulov, Catena Manuscripts of the Greek New Testament: A Catalogue (Text and Studies, Third Series, 25; Piscataway: Gorgias Press, 2021) 54–6.
33 Codex Mosquensis II = Moscow, State Historical Museum, Sinod. Gr. 399, fol. 5. Here it begins with Ἰστέον, ὅτι instead of Καὶ γὰρ as in Fragment IX c.
34 Diekamp, Hippolytos, LV. In the heat of the iconoclast dispute, in April 836, a letter was sent in the name of the three iconodule Patriarchs, Christophorus I of Alexandria, Job of Antioch, and Basil of Jerusalem, to the iconoclast Byzantine Emperor Theophilos, which would later enjoy a wide circulation. For an extensive discussion of the possibility of the synod in Jerusalem in 836 and the Letter's authenticity, see Juan Signes Codoñer, The Emperor Theophilos and the East, 829–842: Court and Frontier in Byzantium During the Last Phase of Iconoclasm (Birmingham Byzantine and Ottoman Studies 13; Farnham: Ashgate 2014).
35 For a critical edition of the Greek versions, see The Letter of the Three Patriarchs to the Emperor Theophilus and Related Texts (J.A. Munitiz, J. Chrysostomides, E. Harvalia-Crook, and Ch. Dendrinos, ed., Camberley: Porphyrogenitus, 1997); for a critical edition of the ninth-century Slavonic translation with reconstructed Greek text, see Dimitri Afinogenov, Mnogosložnyj Svitok: The Slavonic Letter of the Three Patriarchs to Emperor Theophilios (Monographies ACHCByz 41; Paris: Histoire & civilisation de Byzance, 2014). The Svitok matches the only portion of the Tiranensis-fragment that is not preserved in any Greek version, thus confirming that the Greek text behind the Svitok is as old as the Tiranensis-fragment and antedates the prototype behind the Greek versions (Afinogenov, Svitok, 10–14).
36 Hengel, ‘Enstehungszeit’, 9,
37 In Codex Cyprius, the type is found in Matthew (81r), Mark (131v) and Luke (204v).
38 Theophylact wrote in his commentary on Matthew: Ἦν δὲ ὁ μὲν Μάρκος ἀκόλουθος καὶ μαθητὴς Πέτρου⋅ ὁ δὲ Λουκᾶς, Παύλου. Ὁ τοίνυν Ματθαῖος, πρῶτος πάντων ἔγραψɛ τὸ Εὐαγγέλιον Ἑβραΐδι φωνῇ, πρὸς τοὺς ἐξ Ἑβραίων πιστɛύσαντας, μɛτὰ ὀκτὼ ἔτη τῆς τοῦ Χριστοῦ ἀναλήψɛως⋅ μɛτέφρασɛ καὶ τοῦτο Ἰωάννης ἀπὸ τῆς Ἑβραΐδος γλώττης ɛἰς τὴν Ἑλληνίδα, ὡς λέγουσι. Μάρκος δέ, μɛτὰ δέκα ἔτη τῆς ἀναλήψɛως ἔγραψɛ, παρὰ τοῦ Πέτρου διδαχθɛίς. Λουκᾶς δέ, μɛτὰ πɛντɛκαίδɛκα. Ἰωάννης δὲ ὁ θɛολογικώτατος, μɛτὰ τριάκοντα δύο (Migne PG 123:145); and in the commentary on Mark: Τὸ κατὰ Μάρκον Εὐαγγέλιον, μɛτὰ δέκα ἕτη τῆς τοῦ Χριστοῦ ἀναλήψɛως συνɛγράφη ἐν Ῥώμῃ⋅ ἦν μὲν γὰρ οὗτος ὁ Μάρκος Πέτρου μαθητής (Migne PG 123:492).
39 Euthymius wrote in his commentary on the four Gospels: παρ’ οὗ καὶ ὅλον τὸν τοῦ Εὐαγγɛλίου λόγον ὁ Μάρκος μɛμάθηκɛν. Ἔπɛιτα συνɛγράψατο τὸ Εὐαγγέλιον, ὡς μὲν ἱστορɛῖ καὶ Κλήμης ὁ Στρωματɛύς, ἐν αὐτῇ τῇ Ῥώμῃ, κατὰ δὲ τὸν Χρυσόστομον, ἐν Αἰγύπτῳ, παρακληθɛὶς ὑπὸ τῶν ἐκɛῖ πιστῶν, μɛτὰ ἔτη δέκα τῆς τοῦ Σωτῆρος ἀναλήψɛως (Migne PG 129:769).
40 Thus, the scholion from the margin of folio 86r of Oxford, Bodl. Libr. Laud Gr. 33 (GA 50), cited by J. A. Cramer, reads: Τὸ κατὰ Μάρκον Εὐαγγέλιον συνɛγράφη ἀπὸ [MS: ὑπὸ] Μάρκου μɛτὰ δέκα χρόνους τῆς Χριστοῦ ἀναλήψɛως⋅ οὗτος ὑπὸ Πέτρου Ἀποστόλου ἤκουσɛ τὸ Εὐαγγέλιον καὶ συνɛγράψατο αὐτὸ, μαθητὴς ὢν Πέτρου. See J.A. Cramer, ed., Catenae Graecorum patrum in Novum Testamentum (Vol. 1; Oxford: Oxford University Press, 1840) 264.
41 Greek text in Theodor Schermann, ‘Appendices ad indices apostolorum discipulorumque’, Prophetarum vitae fabulosae: Indices apostolorum discipulorumque domini Dorotheo, Epiphanio, Hippolyto aliisque vindicate (Bibliotheca scriptorum Graecorum et Romanorum Teubneriana; Leipzig: B.G. Teubner, 1907) 126–7: Δɛῖ γινώσκɛιν τὸ πῶς συνɛγράφησαν τὰ δ´ ɛὐαγγέλια. Ὅτι μɛτὰ τὴν ἀνάληψιν τοῦ κυρίου ἡμῶν Ἰησοῦ Χριστοῦ τὸ κατὰ Ματθαῖον μɛτὰ ἔτη η´, τὸ κατὰ Μάρκον μɛτὰ ἔτη ιβ´, τὸ κατὰ Ἰωάννην μɛτὰ ἔτη λ´ ἐπὶ Δομɛτιανοῦ τοῦ βασιλέως, τὸ κατὰ Λουκᾶν μɛτὰ ἔτη ιɛ´. The text is also attested with some variation in Vat. gr. 1974 (12th cent.).
42 P. E. Pusey and G.H. Gwilliam, ed., Tetraeuangelium Sanctum (Oxford: Clarendon, 1901) 195–6 (Matthew); 314–15 (Mark); 479–80 (Luke); 604–6 (John). For example, this type of subscription is attested with slight variation in the Peshitta manuscript British Library, Codex Add. 14459 (fols. 1–66), dated to the second half of the fifth century (only extant in Matthew and Mark). For a detailed description, see G. H. Gwilliam, ‘An Account of a Syriac Biblical Manuscript of the Fifth Century with Special Reference to Its Bearing on the Text of the Syriac Version of the Gospels’, Studia Biblica 1 (1885) 151–71. Syriac text and English translation of the subscriptions on p. 157.
43 F. C. Burkitt, Evangelion da-Mepharreshe (2 vols. Cambridge: Cambridge University Press, 1904) 2:160–5.
44 Greek text in Diekamp, Hippolytos, 40: Ἰστέον, ὅτι τὸ κατὰ Ματθαῖον ἅγιον ɛὐαγγέλιον ἐγράφη ɛἰς Παλαιστίνην μɛτὰ ὀκτὼ ἔτη τῆς ἀναλήψɛως ἑβραϊστὶ. τὸ κατὰ Μάρκον ɛὐαγγέλιον ἐγράφη ῥωμαϊστὶ ἐν Ῥώμῃ μɛτὰ δώδɛκα ἔτη τῆς ἀναλήψɛως. τὸ κατὰ Λουκᾶν ɛὐαγγέλιον ἐγράφη ἑλληνιστὶ ɛἰς Ἀλɛξάνδρɛιαν τὴν μɛγάλην μɛτὰ δɛκα-πέντɛ ἔτη τῆς ἀναλήψɛως. τὸ κατὰ Ἰωάννην ɛὐαγγέλιον ἐγράφη ἑλληνιστὶ ἐν Πάτμῳ νήσῳ μɛτὰ τριάκοντα δύο ἔτη τῆς ἀναλήψɛως. The now lost Turin manuscript (Turin, BL, B.VI.25) did not contain Hippolytus’ Chronicle, but rather the Philosophumena attributed to Hippolytus of Rome. See Johann Albert Fabricius, ed., S. Hippolyti episcopi et martyris Opera (2 vols; Hamburg: Liebezeit, 1716), 1, appendix p. 50. The secondary nature of the fragment is further seen in that it connects the Gospel of John to Patmos and omits reference to the reign of Domitian.
45 Greek text and English translation in Stephen Carlson, Papias of Hierapolis Exposition of Dominical Oracles: The Fragments, Testimonia, and Reception of a Second-Century Commentator (OECT; Oxford: Oxford University Press, 2021) 116–17 (Papias F5 apud T5 Eusebius, Hist. eccl. 3.39.16).
46 Irenaeus, Adv. Haer. 3.1.1 (SC 211, 20–5). There is a consensus that Irenaeus’ source is Papias (cf. Eusebius, Hist. eccl. 3.39.15–16). Eusebius confirms the dependency of Irenaeus on Papias regarding the theory of a millennial reign (Hist. eccl. 3.39.11–13). Carlson, Papias, 310–11, labels this passage as ‘Potential Use’ item Y7 of Papias, stating, ‘Since we know that Irenaeus had read Papias, we can see in here his reception of Papias, F4 on Mark, F5 on Matthew, and F6 on traditions from the elders, in combination with other traditional material’ (p. 311 n. Y7).
47 See Carlson, Papias, 300–3 (X12: Eusebius, Hist. Eccl. 3.24.6) and 318–19 (Y15: Origen in Matt. apud Eusebius, Hist. Eccl. 6.25.4–6).
48 In this connection it is perhaps worth noting that in Codex Cyprius (K 017), the oldest witness to Type 3, a geographical modifier, Jerusalem, is only included for the Gospel of Matthew, ɛξɛδοθη υπ αυτου ɛν ιɛροσολυμοις μɛτα χρονους η̅ της του χυ αναληψɛως (017-2), ‘. . . published by him in Jerusalem eight years after the ascension’.
49 Louis Leloir, Saint Éphrem, Commentaire de l’Évangile concordant, texte syriaque (Manuscrit Chester Beatty 709) (Chester Beatty Monographs 8; Dublin: Hodges Figgis, 1963).
50 Louis Leloir, Saint Éphrem. Commentaire de l’Évangile concordant, version arménienne (CSCO 145, Scriptores Armeniaci 2; Leuven: L. Durbecq, 1954) 247 (Latin translation); for discussion of the variants, see Louis Leloir, ‘L'Original syriaque du commentaire de S. Éphrem sur le Diatessaron’, Biblica 40 (1959) 959–70, 965–6. Leloir ascribed the text reflected in the two extant twelfth-century Armenian manuscripts to the fifth century (Saint Éphrem, i).
51 Matthew R. Crawford, ‘The Fourfold Gospel in the Writings of Ephrem the Syrian’, Hugoye 18.1 (2015) 9–51, 24.
52 Gregory of Nazianzus, Carmina dogmatica I.12.6–9 (from the poem de veris scripturae libris). Greek text in Migne PG 37:472–4.
53 For the most recent and extensive study of this family of manuscripts, see Didier Lafleur, La Famille 13 dans l’évangile de Marc (NTTSD 41; Leiden: Brill, 2013). Lafleur did not include 873. See, however, Yvonne Burns, ‘A Newly Discovered Family 13 Manuscript and the Ferrar Lection System’, Studia Patristica 17.1 (1982) 278–89. Lafleur La Famille 13, 365–6, wrongly claims that 13, 346, 543 and 983 omit a subscription (in contrast to his own apparatus on p. 361).
54 We want to thank Martina Vercesi who brought the subscription to Mark in this manuscript and in GA 684 (see below) to our attention in her paper ‘Gospel's Paratexts: Unexplored Avenues in New Testament Manuscripts Research’ at the SBL Annual Meeting in Denver, 2022.
55 For a survey of the Jerusalem colophon, see Wasserman, ‘Greek New Testament Manuscripts’, 86–92.
56 Wasserman, ‘Greek New Testament Manuscripts’, 91.
57 A.1 is written in dactylic hexameter.
58 The Database of Byzantine Book Epigrams (DBBE) is maintained by the Faculty of Arts and Philosophy at Ghent University, see https://www.dbbe.ugent.be/. We have notified the project team of the unrecorded epigram. Moreover, we have noted another unrecorded epigram following the subscription to Mark in GA 684 (Oxford, Bodleian Library MS Holkham Gr. 64), which was not included in our survey, ἀπὸ στόματος ἐκδιδαχθɛὶς τοῦ πέτρου.
59 Cf. Scherbenske, Eric W., Canonizing Paul: Ancient Editorial Practice and the Corpus Paulinum (Oxford: Oxford University Press, 2013) 127CrossRefGoogle Scholar, ‘The sources deemed authentic were used to reconstruct Paul's life, which in turn authenticated them as sources for readers of this edition. In a related manner, these extracanonical traditions supplied a proto-orthodox metanarrative legitimating their own claims of apostolicity and orthodoxy. The inclusion of such traditions in paratexts (prologues, subscriptions, etc.) even ensured their transmission as part of the very scripture they sought to authenticate.’
60 Metzger, Bruce M., ‘Names for the Nameless in the New Testament: A Study in the Growth of Christian Tradition’, Kyriakon: Festschrift Johannes Quasten (ed. Granfield, Patrick and Jungmann, Josef Andreas; vol. 1; Münster: Aschendorff, 1970) 79–99Google Scholar.
61 Metzger, ‘Names’, 79.
62 Nicklas, Tobias, Canon, ‘New Testament and Christian, Ancient “Landscapes of Memory”’, EC 7 (2016) 5–23Google Scholar, 23. Nicklas’ concepts of ‘landscapes’ and ‘sites of memory’ depend on Maurice Halbwachs’ theories of ‘collective memories’ and ‘social frameworks of memories’. See Halbwachs, Maurice, Les cadres sociaux de la mémoire (Paris: F. Alcan, 1925)Google Scholar; and idem, La mémoire collective (Paris; Presses Univ. de France, 1950).