Hostname: page-component-74656f9884-j7cmm Total loading time: 0 Render date: 2026-11-09T09:43:39.367Z Has data issue: false hasContentIssue false
  • English
  • Français

Intertexts in the Gospel of Matthew

Published online by Cambridge University Press:  30 April 2004

Ulrich Luz
Ulrich Luz
Affiliation:
University of Bern
Rights & Permissions [Opens in a new window]

Abstract

An abstract is not available for this content. As you have access to this content, full HTML content is provided on this page. A PDF of this content is also available in through the ‘Save PDF’ action button.

Information

Type
Research Article
Information
Harvard Theological Review , Volume 97 , Issue 2 , April 2004 , pp. 119 - 137
Copyright
© 2004 Cambridge University Press

The introduction of the concept of intertextuality into critical discourse in the late 1960s transformed the discussion of all kinds of literature, including biblical literature. The relationship between the texts produced by early Christians and “the Bible”—that is, the Septuagint—had, of course, occupied biblical scholars for centuries. But the explicit formulation of intertextuality as a new concept, and the development of a vocabulary for describing and discussing in detail its operation, have made it possible to bring greater refinement and precision to the study of the relationship between “the Bible” and the texts that would become the “New Testament.” In this essay, I review some of the ways in which the concept of intertextuality has been formulated and modified, and I introduce as well some terminology developed by other scholars to articulate the concept. I then use these theoretical tools to re-examine the relationship between the Gospel of Matthew and some of its intertexts, including the Bible.

Intertexts and Intertextuality

What do I mean when I speak about “intertexts”? Julia Kristeva, whose influential studies are the matrix of all discussions of intertextuality, stated: “Every text constructs itself as a mosaic of quotations; every text is an absorption and transformation of another text: The notion of intertextuality takes the place of the notion of intersubjectivity.”1

“Tout texte se construit comme mosaïque de citations; tout texte est absorption et transformation d'un autre texte: A la place de la notion d'intersubjectivité s'installe celle d'intertextualité.” Julia Kristeva, Sémeiotiké. Recherches pour une sémanalyse (Paris: Seuil, 1969) 146.

Moving toward a definition, Kristeva wrote: “Let us give the name ‘intertextuality’ to the textual interaction which takes place within a single text. For the subject, intertextuality is a notion which indicates how a text reads history and inserts itself into history.”2

“Nous appellerons intertextualité cette inter-action textuelle qui se produit à l'intérieur d'un seul texte. Pour le sujet connaissant, l'intertextualité est une notion qui sera l'indice de la façon dont un text lit l'histoire et s'insère en elle.” Julia Kristeva, “Narration et transformation,” Semeiotica 1 (1969) 443.

History and society are reflected in texts and can themselves be read textually.3

Ibid.

In other words, “Every text appeals to the reader's memory of other texts.”4

“Tout texte appelle à la mémoire du lecteur, de la lectrice d'autres textes.” Daniel Marguerat and Adrian Curtis, “Préface,” in Intertextualités. La Bible en échos (ed. idem; MdB 40; Geneva: Labor et Fides, 2000) 5.

Seen in this way, every text becomes an “intertext,” as Roland Barthes says:

Every text is an intertext; other texts are present in it, on different levels, in more or less recognizable forms: the texts of the earlier culture and those of the surrounding culture. Each text is a new tissue of past quotations … a general field of anonymous formulas, whose origin is only rarely detectable; of unconscious or automatic quotations, reproduced without quotation marks.5

“Tout texte est un intertexte; d'autres textes sont présents en lui, à des niveaux variables, sous des formes plus ou moins reconnaissables: les textes de la culture antérieure et ceux de la culture environnante; tout texte est un tissu nouveau de citations révolues … un champ général de formules anonymes, dont l'origine est rarement repérable, de citations inconscientes ou automa-tiques, données sans guillemets.” Roland Barthes, “Texte,” in Encyclopaedia Universalis (32 vols.; Paris: Encyclopaedia Universalis, 1990) 22:370–74, at 372.

Intertextuality is nothing less than the textual shape of how culture, history, and society are engraved in texts. This concept transcends a text-immanent structuralism and shows how texts are mirrors or echoes of the world.6

Roland Barthes (Roland Barthes par Roland Barthes [Paris: du Seuil, 1975] 78) uses the phrase “une chambre d'échos” (“a chamber of echoes”).

“Intertextuality thus becomes less a name for a work's relation to particular prior texts than a designation of its participation in the discursive space of a culture.”7

Jonathan Culler, “Presupposition and Intertextuality,” in idem, The Pursuit of Signs (London: Routledge & Kegan Paul, 1981) 103.

Intertextuality is a comprehensive model of textuality, and it is possible to accentuate certain of its features so that the theory may be applied in very different ways. Intertextuality can be formulated as a synchronic principle describing the structure of texts: “Other texts” are present in a given text; they are assembled within it and form part of its structure. Intertextuality can also be formulated with a stress on the diachronic dimension of textual analysis: Intertexts are memories preserved by a text—for example, sources, reminiscences, models, or patterns. Intertexts can be specific or general: It is possible to narrow the definition of “intertexts” to specific texts that can be identified, and it is possible to widen the definition, so that all texts of a specific culture or a specific time are present in a text, regardless of whether an author or reader is conscious of their presence.8

See, e.g., Harold Bloom, A Map of Misreading (New York: Oxford University Press, 1975) 3: “There are no texts, but only relationships between texts.”

Intertextuality can be a tool for decentering the author, whereby the author is reduced to the status of a filter through which passes an endless stream of voices, echoes, texts, and reminiscences.9

Manfred Pfister, “Konzepte der Intertextualität,” in Intertextualität. Formen, Funktionen. Anglistische Fallstudien (ed. Ulrich Broich and Manfred Pfister; Tübingen: Max Niemeyer, 1985) 21: An author is “a ‘chamber of echoes,’ full of the sound and the roar of other texts” (“eine ‘Echo-kammer,’ erfüllt vom Hall und Rauschen fremder Texte”).

It can be a tool for the deconstruction of texts, whereby the text becomes something like a cloud of music without precise meaning, into which the intended sense of a text may disappear completely; or, in the words of Roland Barthes, the text becomes “a music of figures, of metaphors, of thought-words; it is the signifier as a siren.”10

Barthes, Roland Barthes, 148: “une musique de figures, de métaphores, de pensées-mots; c'est le signifiant comme sirène.”

But the concept of intertextuality can also be used as a tool for the reconstruction of the sense of a text, for the recovery of the author's context with the goal of differentiating between the author's own voice and the voices of his sources and reminiscences. And it can be a tool for formulating the sense of a text in a more precise way, as when one clarifies the affinities and differences between a text and its various intertexts. In this case, the reader does not passively absorb an ambient cloud of music, but rather listens attentively to discern single tones and melodies.

Intertexts may be the product of the text or of the reader. In the first case, they belong to the rhetorical strategy of a text; in the second case, their detection is a performance by the text's readers.11

This is the approach of Michael Riffaterre (La syllepse intertextuelle [Poétique no. 40, 1979] 496): “L'intertextualité est un mode de perception du texte … le méchanisme propre de la lecture” (“intertextuality is a mode of perceiving the text … [It is] the proper mechanism of reading”). Riffaterre (ibid.) makes a distinction between this “literary reading” of a text which creates the “significance” of a text and a merely “linear reading” which recovers its “sense” only.

If they are the product of the text, the number of intertexts in a given text is limited, because a text does not have an unlimited num-ber of strategies, and an author has neither unlimited knowledge nor an unlimited number of intentions. Such intertexts are always temporally prior to the text. If they are the product of the reader, however, the number of intertexts is unlimited: Every reader has both the capability and the liberty to discover new intertexts in a given text; moreover, these intertexts can be temporally either prior or posterior to the given text. Intertextuality on the level of the text is primarily descriptive; it facilitates the precise description of the strategies of a text. A reader-oriented concept of intertextuality, however, tends to widen the meaning of a text. An example of the limited widening of meaning is the intertextual reading of a biblical text according to the principles of classical church exegesis or of fundamentalistic exegesis, where only other canonical texts may be cited as intertexts. An unlimited widening of meaning is proposed by Riffaterre: “The intertext is the reader's perception of the relations between an œuvre and others that either precede or follow it.”12

“L'intertexte est la perception, par le lecteur, de rapports entre une œuvre et d'autres qui l'ont précédée ou suivie” (ibid.).

In short, the concept of intertextuality is very fluid and can be used in many different ways.

In light of this conceptual fluidity, I want to qualify my own interest in intertextuality in three ways. First, as an exegete, I am interested in a model of intertextuality that is text-oriented, i.e., author-oriented. My concern is to seek specific and identifiable intertexts in a manner that is subject to control or verification. Consequently, I focus on intertexts that are consciously invoked by an author and that are part of the rhetorical strategy of a text. But I do not want to neglect or deny the existence of intertexts that are not intentionally invoked and that are not exegetically verifiable. Second, as an exegete and historian, I am interested in a model of intertextuality that is attentive to history and interprets specific intertexts as reflections of a specific historical and cultural situation. Hence, I am not prepared to give up the quest for the real, extratextual author of a text. And I am not an advocate of an exclusively synchronic and text-immanent model of intertextuality. But I do concede that the quest for intertexts is primarily a synchronic quest, and that the synchronic quest for the presence and function of specific intertexts in a given text always comes first. Third, as a hermeneut, I do not want to neglect the importance of reader-oriented concepts of intertextuality, nor to deny the fact that every text is fundamentally intertextual. But readers are always concrete persons in concrete situations. They are among the very first readers of a text, or they are readers whom we can locate in the reception history of a text, or they are readers of today. Each reads the texts in his or her own way, utilizing the specific intertexts which are important to that individual reader. My hermeneutical model, therefore, is not the model of Roland Barthes—the text as a diffuse cloud of music generated by “the signifier as a siren”—but rather the model of a musical score, which makes it possible that one and the same piece of music can be interpreted by different musicians in different ways. The task of the exegete is not to become lost in a diffuse cloud of music, but to compare the different performances of a specific piece of music, including his or her own, with the musical score and the instructions of the composer encoded within it.

A Short Lexicon of Intertextuality

In this essay I will foreground my first and second interests, as exegete and exegete-historian. Here I describe two conceptual and methodological refinements of the concept of intertexts, introducing a number of terms that will prove useful in the following discussion of Matthew.

The first is Gérard Genette's model of intertextuality, clearly summarized in the beginning of his book Palimpsestes.13

Gérard Genette, Palimpsestes. La littérature au second degré; (Paris: Seuil, 1982) 7–16.

Genette defines “intertextuality” in the narrower sense of the word as “a relation of ‘co-presence’ between two or more texts … in most cases, taking the form of the effective presence of one text in another.”14

Ibid., 8: “une relation de coprésence entre deux ou plusieurs textes … le plus souvent, par la présence effective d'un texte dans un autre.”

This includes phenomena ranging from quotation to allusion to plagiarism. Genette makes a distinction between “intertexts” in this sense and “paratexts.” Paratexts include titles, prefaces, marginal glosses, and footnotes.15

Ibid., 10–11.

Different from these intertexts are “metatexts.” In Genette's terminology, a metatext is a commentary, which explains its pretext while preserving a critical distance between itself and its pretext.16

Ibid., 11–12. In this essay, however, I will not use the term in this sense; rather, I will use “metatext” in its more general sense, namely, as a later text containing quotations or allusions to an earlier text, a “pretext.”

In distinction from the relationship between metatext and pretext, Genette defines the relationship between a “hypertext” and its “hypotext.” A hypertext is a secondary text that is written entirely on the basis of a preceding pretext, the hypotext, but without being a formal commentary on its hypotext. For example, Virgil's Aeneid is a hypertext to Homer's Odyssey. Finally, there is the “architext”: for Genette, this term designates the general type or model underlying a specific text. It is the equivalent of the traditional term “genre.”

The second model I want to introduce is that proposed by Manfred Pfister, a German scholar of English literature. Pfister aims to distinguish between the different ways that pretexts function within their metatexts, and also to establish a terminology for measuring the relative intensity of a pretext's presence in its metatext. To this end he defines six categories:17

Pfister, “Konzepte,” 25–30.

Referentiality. A pretext may simply be mentioned by a metatext, e.g., as a decorative element or as a mere reference. Alternatively, the pretext's content and context may be fully exploited, as in the case of a quotation, when the literary context and the intention of the pretext is fully considered by the metatext. Accordingly, the level of referentiality of a pretext in relation to its metatext may be either low or high.

Communicability. Pfister defines this as the degree of consciousness that an intertextual reference implies on the part of an author or reader. Allusions or idioms that belong to a common cultural heritage are often used unconsciously and thus have a low degree of communicability. Specific pretexts that are explicitly marked in a metatext, or that are consciously concealed by a plagiarist, have a high degree of communicability.

Autoreflexivity. Autoreflexivity is high when a pretext is not only mentioned in a metatext, but explicitly reflected within it.18

I will not make use of this category, however, because it seems to me almost indistinguishable from “communicability.”

Structurality. The intensity of structurality is particularly high when “a pretext [serves as] the structural pattern of an entire text.”19

Pfister, “Konzepte,” 28: “ein Prätext [wird] zur strukturellen Folie eines ganzen Textes.”

For example, the dependency of the Epistle to the Ephesians upon the Epistle to the Colossians displays a high degree of structurality. (In Genette's terms, Colossians is the hypotext upon which Ephesians, as hypertext, depends.)

Selectivity. Selectivity of intertexts is more intensive in specific and pointedly marked intertexts than in mere topoi or motifs that are appropriated without being explicitly marked.

Dialogicity. The degree of dialogicity of an intertext is particularly high when there is a tension between pretext and metatext, so that there is an explicit dialogue between the two texts.

Pfister's terminology is especially useful insofar as it allows for a greater degree of specificity when describing both the quality and degree of intertextuality than that permitted by the terms commonly used to classify intertextual relationships, such as “quotation,” “allusion,” or “motif.”

Intertexts in the Gospel of Matthew: Introductory Remarks

Even if I use the term “intertexts” in the relatively narrow sense of “specific intertexts,” and even if I put my primary emphasis on pretexts that were intentionally invoked by the evangelist and that were possibly known to his first readers as well, my quest for intertexts nevertheless covers a broader territory than that defined by the traditional investigation in the use of the Bible in the Gospel of Matthew.20

George W. Buchanan (Introduction to Intertextuality [Lewiston, N.Y.: Mellen, 1994]) unfortunately limits his program for intertextual exegesis to the Hebrew Bible.

This is true in three respects. First, not only the Bible, but also the Gospel of Mark, the Sayings Source (Q), and possibly other texts related to the Jesus tradition are intertexts. One problem is immediately evident: Matthew has used his two main groups of intertexts, the Bible and his sources about Jesus, in very different ways. Second, not only those specific texts which are quoted, alluded to, or used in the Gospel of Matthew are intertexts. While searching for intertexts, we must look also for hypotexts (to use Genette's term) that shape the structure of the Gospel as a whole; for other structuring elements that can be connected with specific intertexts; and for motifs, persons, or historical events that are related to specific pretexts. And third, because the use of intertexts is part of the rhetorical strategy of a text, we must consider the reception of intertexts by its implicit reader, as well as by the actual first readers of that text.

The Gospel of Mark as Intertext

In the terminology of Genette, the Gospel of Mark is a hypotext that Matthew has absorbed and transformed in his hypertext.21

For the different possibilities of “transposition,” or transformation, of pretexts in a metatext, see Genette, Palimpsestes, 291–92.

When we apply the categories introduced by Pfister to measure the relative intensity of intertextual presence, the result is interesting. The Matthean metatext displays a very high degree of structurality in relation to the Gospel of Mark; indeed, the Gospel of Mark determines the whole structure of Matthew's Gospel. I would thus describe the Gospel of Matthew as an enlarged “new edition” of the Gospel of Mark. Matthew does not have an architext, because the genre of “gospel” did not yet exist when it was written; it has only a specific hypotext, the Gospel of Mark. Matthew also displays a high degree of referentiality. Matthew absorbs not only the plot and structure of Mark, but also its basic theological concerns, such as the rejection of Jesus in Israel, the mission to the Gentiles, the cross, and the role of suffering in discipleship. Matthew's Gospel is also a transformation of Mark's Gospel. In Genette's terms, the two most important methods of transformation are “extension” (by adding materials) and “concision” (by shortening and condensing individual narratives).22

Ibid., 364–72, 331–40.

The consequence of all this is that there is also a high degree of communicability between the two texts, so that it is possible to interpret the Gospel of Matthew as a hypertext engaged in a conscious dialogue with its hypotext.

It is all the more astonishing, then, that Matthew does not make this dialogue explicit; the degrees of selectivity and dialogicity are low. The Gospel of Matthew never mentions its hypotext; it never thematizes or problematizes its relation to it. Remarkably, we find no explicit quotation of the all-determining hypotext, and this curious absence deserves comment. Matthew certainly knew the Gospel of Mark, but did Matthew's audience know it?

A conjectural answer to this question may be supplied on the basis of evidence external to the text. The Gospel of Mark, in my opinion written in Rome, became a rather widespread and—in the Christian congregations—well-known book within a few decades. There is no reason to reject the thesis that at least some of Matthew's first readers, or hearers, must have known it, not to mention those hearers and readers in other churches where the Gospel of Matthew became known very quickly. But it is surprising that the evidence within the text amounts to almost nothing. It cannot be proved that the Gospel of Matthew presupposes the knowledge of the Gospel of Mark by its implicit reader. Only in a very few cases has Matthew omitted stories narrated in Mark. But in no case, not even in the case of Matthew's omission of Mark 4:26–29 (the parable of the seed that grows without human aid), is it necessary to assume that the reader knows what Matthew has omitted in order to understand his text.23

Compare Matt 13:1–23 to Mark 4:1–20, 26–29. In the rare cases where Matthew has omitted Markan texts, the reason for the omission is either compositional (e.g., in the case of Mark 1:23–27 and Mark 12:41–44) or motivated by the desire to eliminate doublets (as in the case of Mark 13:33–37).

Drastic abbreviations of Markan narratives in the Gospel of Matthew are never so curt that the Matthean version of a story is understandable only by those readers or hearers who know the Markan pretext.24

This is also true for Matt 8:28–34//Mark 5:1–20 and for Matt 14:3–12//Mark 6:17–29.

A possible exception is Matthew's account of the healing of the paralytic man (Matt 9:2–8), which omits Mark's detail (2:4) that the friends of the paralytic made an opening in the roof through which they lowered his stretcher. In Mark, this notice is followed by the remark that “Jesus saw their faith” (v. 5). But Matthew's statement that “Jesus saw their faith” is understandable without this detail; their faith is witnessed by their act of carrying the paralytic on his stretcher to Jesus.

Matthew's treatment of the Bible is entirely different from his treatment of Mark. Why? Several considerations may be helpful. First, the Bible is a canonical text of special dignity for Matthew, unlike the Gospel of Mark. It is characteristic of early Christianity in the first and second centuries that only the Bible is quoted as Scripture, whereas the acknowledgment of Christian intertexts takes different forms, even when the Gospels are quoted.25

Wolf D. Köhler, Die Rezeption des Matthäusevangeliums in der Zeit vor Irenäus (WUNT II/24; Tübingen: Mohr-Siebeck, 1987) 518–19.

Second, it is generally true in the ancient world that the way in which pretexts are used depends on their status and authority. Thus, classics and well-known authors are explicitly quoted more frequently than unknown or contemporary authors.26

Loveday Alexander, “L'intertextualité et la question des lecteurs. Réflexions sur l'usage de la Bible dans les Actes des apôtres,” in Marguerat and Curtis, eds., Intertextualités, 204–7. Later Greek historians prefer to quote the classics (e.g., Thucydides, Herodotus, and Xenophon) rather than contemporary authors.

Third, the way in which pretexts are invoked—that is, whether they are quoted or not quoted—depends also on the genre of the metatext. Deliberative and judicial genres and commentaries tend to quote frequently; in narrative genres, including Greek historiography, quotations are rather rare.27

See Dieter A. Koch, Die Schrift als Zeuge des Evangeliums (BHT 69, Tübingen: Mohr-Siebeck, 1986) 11 n. 3. The Gospel of Luke, which belongs to the genre of historical monographs, must have been a rather strange book for its pagan readers precisely because it includes numerous quotations from only one “classical” book, the Septuagint.

A close analogy in genre to the Gospel of Matthew is provided by those Jewish texts that Geza Vermes and others call “parabiblical texts”28

Geza Vermes, Scripture and Tradition in Judaism (2d ed.; StPB 4; Leiden: Brill, 1983) 67–126; and Florentino García Martinez, The Dead Sea Scrolls Translated (Leiden: Brill, 1994) 217–99; see also Ulrich Luz, “Das Matthäusevangelium—eine neue oder eine neu redigierte Jesusgeschichte,” in Biblischer Text und theologische Theoriebildung (ed. S. Chapman et al.; Biblisch-Theologische Studien 44; Neukirchen: Neukirchener, 2001) 68–71.

or “rewritten Bibles.” These are texts that narrate the foundational history of the Bible anew, taking into account the needs of the present. Among them are Jubilees, the Liber antiquitatum biblicarum of Pseudo-Philo, the Genesis Apocryphon, and other Qumran texts. From a wider point of view, biblical texts like the Priestly work or Chronicles should be mentioned here. They, too, do not explicitly quote their biblical intertexts, but they do retell their foundational history. It seems to me that Matthew did retell his own foundational story—namely, the story of Jesus as narrated in the Markan Gospel—in an analogous way.

The Sayings Source (Q) as Intertext

Matthew did not use the text of the Sayings Source (Q) as a structuring hypotext; rather, he disrupted its compositional integrity and inserted its individual elements into new contexts. Moreover, only in some sections of his Gospel did Matthew insert blocks of material from Q.

29

Q 6:20–49 for Matt 5–7; Q 7:18–35 for Matt 11:2–19; Q 9:57–10:24 for Matt 9:37–10:40; Q 11:14–32 for Matt 12:22–35; Q 11:39–52 for Matt 23:1–36. See Ulrich Luz, “Matthäus und Q,” in Vom Jesus zum Christus (FS P. Hoffmann; BZNW 93; Berlin: de Gruyter, 1998) 208–12.

Therefore, not only selectivity and dialogicity but also structurality with regard to Q are weak in Matthew. For Matthew, Q did not have its own literary dignity as a compositional unity; the Sayings Source was a mere collection of materials that he freely excerpted. Matthew even gave preference to the Gospel of Mark when positioning the five discourses within his Gospel, and he respected the order of the Sayings Source almost only insofar as it corresponded to the order of Mark. In four of Matthew's five discourses, the materials from Mark precede those from Q. All this shows that Matthew's esteem for Q as a literary text was very different from his esteem for Mark's Gospel. The specific dignity Q had for Matthew is evident insofar as he excerpted its words of Jesus almost without omission and preserved their wording rather more faithfully than Luke did. To summarize, Matthew inserted the sayings of Jesus found in Q into his own story of Jesus, which he narrated on the basis of the Markan hypotext. Thus, Q became superfluous as a pretext. That the Sayings Source was not preserved by the church as an independent document is a fact that Matthew probably would not have resented very much. For him, Q was something like the

of the apostles for Justin: a written codification of the living words of the Lord Jesus.

The Bible as Intertext

I want to explore five different ways in which the Gospel of Matthew engages biblical intertexts. Beginning with cases where Matthew conjures biblical texts by naming persons and places of the Bible, I then examine how the title of the Gospel suggests the title of the first book of the Pentateuch. Next, I discuss some criteria for describing and evaluating allusions, both intentional and unintentional. Finally, I consider two forms of biblical quotation in Matthew: simple quotations of biblical texts, and the formulaic fulfillment quotations.

Reminders of biblical stories, persons, and places; the genealogy

The Gospel of Matthew is full of the names of biblical persons and places that would have been well known to its audience. These names function as abbreviations which call to mind various biblical texts, so that readers can enlarge upon Matthew's terse mentions by drawing on their knowledge of the Bible. Examples include Abraham (3:8); Solomon (6:29); Tyre and Sidon, and Sodom and Gomorrah (11:21–24); the Ninevites and the queen of the South (12:41–42); and Abel and Zechariah (23:35). Of particular interest is the first major section of the Gospel, the genealogy (l:2–17), because it reveals something specifically Matthean: Matthew narrates here the history of Israel in the condensed form of a genealogy, which is fully comprehensible to its audience only when their treasury of biblical knowledge is opened. This is particularly true of the four women mentioned in the genealogy. In the beginning of his Gospel, Matthew has condensed the long history of Israel into a genealogy that ends with the genesis of Jesus, the Messiah (1:18). This short version of the history of Israel acquires a new function: What had been the foundational history of Israel now becomes the prehistory of a new foundational story, the new “book of Genesis” of Jesus Christ (1:1). In the case of the genealogy, the selectivity of the biblical intertext is very low—the genealogy refers to the whole Bible—but its communicability and structurality are high.

The title of the Gospel and its pentateuchal structure

The title of the Gospel (Matt 1:1) refers to the Bible as hypotext. Convincing are the arguments of William D. Davies and Dale C. Allison as well as Moisés Mayordomo-Marín that the title

refers to the whole book of Matthew and not only to its first chapter;

30

William D. Davies and Dale C. Allison, The Gospel according to St. Matthew (3 vols.; ICC; Edinburgh: T&T Clark, 1988–1997) 1:149–55; Moisés Mayordomo-Marín, Den Anfang hören (FRLANT 180; Göttingen: Vandenhoeck, 1998) 208–14.

the word

suggests this.

31

Mayordomo-Marín (Den Anfang, 211–12) also makes a second proposal: In the literature of antiquity it is not necessary that the title of a book summarize its entire content.

probably refers to the Greek name of the first book of the Bible which is already well attested in the first century C.E. Matthew names his text “book of Genesis,” but qualifies it as the new “Genesis of Jesus Christ.” By choosing a biblical name for his book, Matthew makes an implicit claim to biblical authority. Matthew's intention is to tell a foundational story for the people of God, just as the Bible does. But his foundational story is a new story, the story of Jesus Christ. We have in Matthew's use of the Pentateuch a clear case of what Genette would call a hypotext. But this biblical hypotext of the Gospel of Matthew is very different from its primary hypotext, the Gospel of Mark, because Matthew narrates a new story, like Virgil in the Aeneid; the Bible is something like a “secondary hypotext” to our Gospel. The degree of selectivity in Matt 1:1 is rather high, but this is not at all important, because its referentiality is low: Matt 1:1 understands

differently than Gen 2:4 and 5:1. The specific intertexts function here only as a sign of a structural affinity, or, to use Genette's term, as a sign of the role of the biblical book of Genesis, or of the whole Pentateuch, as hypotext.

Another structuring element that recalls the pentateuchal hypotext are the five great discourses of the Gospel of Matthew; thus does Matthew emulate the fivefold division of the Pentateuch. In 5:1–2, and particularly in 7:28–29, Matthew recalls the experience of Moses on Mount Sinai, and these notices are in harmony with the frequent reminiscences of Moses throughout Matthew's Gospel32

Dale C. Allison, The New Moses: A Matthean Typology (Minneapolis: Fortress, 1993). I do not assume, however, that the various allusions to Moses form a coherent christology of Jesus as a “new Moses.”

and with its Immanuel christology (see especially 1:22–23 and 28:20). The promise of the presence of God with his people is a motif that pervades the whole of the Bible; to use a musical analogy, it is a cantus firmus sounding throughout the Bible.33

See Horst D. Preuß, “… ich will mit dir sein!,” ZAW 80 (1968) 139–73; Willem C. van Unnik, “Dominus vobiscum: The Background of a Liturgical Formula,” in New Testament Essays (FS T. W. Manson; ed. A. J. B. Higgins; Manchester: Manchester University Press 1959, 270–305); and David D. Kupp, Matthew's Emmanuel (NTSMS 90; Cambridge: Cambridge University Press, 1996) 138–56.

This cantus firmus, the Immanuel motif, is used by Matthew to interpret Jesus. Its sonority gives Matthew's book the character of a new, Bible-like foundational story that narrates the saving presence of God with his people.

Allusions

I now turn to the use of specific biblical pretexts—first to allusions, and subsequently to quotations. As a source from which allusions and quotations are drawn, the Bible functions as a kind of reference text.

The history of modern biblical exegesis attests that it is extremely difficult for scholars to reach a consensus about what constitutes an intentional biblical allusion. Questions that continue to attract discussion and debate include: 1) Where is the borderline between an intentional allusion to a biblical text and an unconscious biblicism that would come quite naturally to an author deeply influenced by the language of the Septuagint? 2) What is the difference between a reading of the Gospel of Matthew by a Jewish reader or hearer who has been familiar with the Bible from youth, and a reading by a Gentile Christian who is not so well versed in the Bible? That is, to what extent is the implicit reader of the Gospel of Matthew a “biblically informed” reader?

I think that, as a general rule, we should assume that the implicit reader of the Gospel of Matthew possessed a very deep familiarity with the Bible. Matthew, like most of his first readers, was a Jewish adherent of Jesus who knew Greek. Through contact with synagogues and also with churches, whose members included Christian scribes, the author and his implicit audience were familiar with the Bible. Unlike modern “overinformed” readers, however, they were familiar with at most a few books, if not only with the Bible itself. Allusions as a form of intertextuality are consequently very common in both late biblical and early Jewish writings. This can be illustrated, for example, by the numerous biblical allusions in apocalyptic texts; it is also apparent in the operation of the rabbinic principle of

, i.e., the interpretation of one verse in light of another verse containing similar words. Indeed, if two texts have just two characteristic words in common, the exegete can interpret one in light of the other. This hermeneutic principle presupposes an audience that is highly attuned to allusions to biblical texts, even if the textual signals are minimal. Thus, we can assume that the biblical intertext is present to a high extent for the readers whom Matthew had in mind, and that the practice of allusion was not only familiar to but even expected by them.

34

Dale Allison (The Intertextual Jesus: Scripture in Q [Harrisburg, Pa.: Trinity, 2000] 1–24) has argued this point regarding the audience of Q. What he says is valid for Matthew as well.

Richard Hays and Dale Allison are among the scholars who have proposed rules for evaluating whether a similarity between texts may in fact be described as an allusion.35

Richard Hays, Echoes of Scripture in the Letters of Paul (New Haven, Conn.: Yale University Press, 1989) 29–32; and Allison, The New Moses, 10–13.

Following their lead, I will formulate four criteria for identifying biblical allusions in the Gospel of Matthew: 1) A Gospel text and its presumed biblical intertext must share more than one of the following elements: specific lexical items, word order, syntax, themes, images, or structure. 2) The biblical intertext should have been recognized as such by earlier readers; that is, it should have a pedigree in the history of interpretation. 3) The probability of allusion is higher if the presumed biblical intertext is used elsewhere by the author, or if it is taken from a biblical book that is often quoted by the author. 4) The probability of allusion is higher if the presumed biblical allusion is in harmony with a coherent interpretation of the whole text in which it appears. I would now like to apply these criteria in a discussion of two narrative texts in Matthew in which biblical allusions have been spotted: 2:13–23, the flight of the holy family to Egypt and their return to Israel; and 28:16–20, the final verses of the Gospel. For each text, I will consider four categories of biblical references: 1) explicit quotations; 2) allusions intended by the author; 3) unintentional allusions that may be “discovered” by the reader; and 4) “biblical” language and themes.

Matt 2:13–23. 1) This passage contains three explicit quotations, each marked by a fulfillment formula, in vv. 15, 18, and 23. 2) An obviously intentional allusion to Exod 4:19–20 is found in vv. 19–21. This allusion is obvious, because a) eight words are identical in the pretext and the metatext; b) the plural

is very awkward in the Matthean context; and c) the biblical context of the pretext is the story of the birth of Moses, which is alluded to more than once in Matthew 1–2. 3) In addition to this obviously intentional allusion, exegetes have postulated other allusions whose relation to a biblical pretext, however, is far less obvious. We cannot speak about intentional allusions in these cases; their degree of communicativity is very low. Among the possible pretexts is Gen 46:2–4, the journey of Joseph to Egypt, as proposed by David Daube.

36

See David Daube, The New Testament and Rabbinic Judaism (London: Athlone, 1956) 189–92; for more references see Ulrich Luz, Matthew 1-7 (trans. Wilhelm C. Linss; Minneapolis: Augsburg, 1989) 145 n. 15. George W. Buchanan (The Gospel of Matthew I [Mellen Biblical Commentary, NT 1; Lewiston, N.Y.: Mellen Biblical Press, 1996] 94–95) sees in Matt 2:13–15 a typology of an old and new Joseph.

Another proposed allusion is the trace of the Passover Haggadah in the word

(2:14; compare Exod 12:30–31).

37

Davies and Allison, The Gospel according to St. Matthew, 1:261.

Also, numerous biblical and Jewish traditions identify Egypt as a traditional destination for fugitives.38

Ibid., 259.

In these cases I do not think that we can speak about intentional allusions, either because the texts do not share enough identical words, or because no single biblical text can be specified as the pretext of the allusion. Nevertheless, in each of these cases, the biblical character of the text conjures biblical associations. Readers may, and even should, “discover” biblical associations in these verses; the text encourages readers' creativity, but does not direct it.39

The history of interpretation in both ancient and modern times witnesses the creativity of readers thus set free by the text!

4) Matt 2:13–23 contains several verbal formulas or narrative patterns that have in general a biblical character, such as

(v. 21);

(v. 16);

40

Five times in LXX.

and the pattern of exodus and return that underlies the whole passage.41

Davies and Allison, The Gospel according to St. Matthew, 1:263.

Such formulas and patterns demonstrate the biblical color of Matthew's language and immerse the reader in a “biblical atmosphere” that facilitates associations between the Gospel and numerous biblical texts. But the formulas and patterns are not themselves allusions.

On the narrative level, the story of Matt 2:13–23 is comprehensible to a reader who brings no biblical associations to the text. Consider the indication of time,

, in v. 14: It is not at all surprising that Joseph, who surely had his dream in the night, immediately departs for Egypt to deliver his family from imminent danger, and the reader who does not know Exodus 12 will not be confused. Both the intentional and the “discoverable” biblical allusions, however, endow the text with a kind of biblical depth-dimension. Readers may notice or feel that many possible biblical associations are offered by the text, and behind this network of hermeneutic possibilities they may divine the mystery of the biblical God who is at work in the history of Jesus. By contrast, it is important that readers cannot use the obviously intentional biblical quotations and allusions in order to construct a one-dimensional biblically grounded christology. For example, readers familiar with the Bible will remember that Hos 11:1 speaks of Israel as the son of God who was called from Egypt, while Exod 4:19–20 narrates the return of the family of Moses to Egypt. Such examples show that Matthew does not simply compose his story according to traditional biblical patterns;

42

In Matt 2:13–23, Jesus is not portrayed in a linear way as a “new Moses”; rather, he really is a new Moses!

rather, he tells a new story about the salvation of the promised royal infant by the God of the Bible, a story that is in many respects surprising to its biblically informed readers.

Matt 28:16–20. The final verses of Matthew are quite remarkable for their many intratextual allusions. Many motifs, themes, and verses from the Gospel of Matthew are here taken up again and brought to culmination in the Gospel's conclusion. These intratextual allusions are more important than the equally numerous—but not equally clear—intertextual allusions.43

Ulrich Luz, Das Evangelium nach Matthäus (4 vols.; EKK 1; Neukirchen: Neukirchener; Düsseldorf: Benziger, 1985–2002) 4:436.

Turning to the latter, I again arrange them in four categories. 1) No direct biblical quotations are present in this text. 2) Verse 18b, however, contains a clearly intentional allusion to Dan 7:13–14. The texts share three words, and Dan 7:13–14 is quoted or alluded to in two other places in the Gospel (24:30–31; 26:64). It is striking, however, that Matt 28:18 alludes to these verses in a very free way. Dan 7:13–14 does not refer to the Parousia, as do Matt 24:30–31 and 26:64, but to the present rulership of the exalted Lord. Here the metatext, Matt 28:18, completely dominates its pretext, Dan 7:13–14; its referentiality is very low.

3) All the other allusions spotted in our text by exegetes belong to the third group: namely, allusions that biblically informed readers may “discover,” but that were hardly intended by the author. Among these are the allusions to the edict of Cyrus (2 Chr 26:23) emphasized by Hubert Frankemölle: The structure of the two texts is similar, but the verbal identities are not numerous and do not include specific words.

44

Hubert Frankemölle, Jahwebund und Kirche Christi (NTAbh n.s. 10; Münster: Aschendorf, 1973) 51–53.

Moreover, this pretext is unknown to the rest of the New Testament, and Matthew's readers could hardly have taken from it any hermeneutic guidance.45

This is even more applicable to Gen 45:9–11, a text considered as an important intertext by Bruce Malina, “The Literary Structure and Form of Matt 28:16–20,” NTS 17 (1970–1971) 96.

To this group of unintentional allusions belong also the biblical texts about the pilgrimage of the peoples of the world to Mount Zion at the end of time; these “allusions” are stressed by Terrence Donaldson and Peter Stuhlmacher.46

E.g., Ps 2:6–8; Isa 2:2–4; 25:6; 56:7; Zech 2:10–16; 14:16–19. Compare Terrence L. Donaldson, Jesus on the Mountain (JSNTSup 8; Sheffield: JSOT Press, 1985) 183–87, 197–202; and Peter Stuhlmacher, “Zur missionsgeschichtlichen Bedeutung von Mt 28,16-20,” EvT 59 (1999) 115–17. For a critique, see Luz, Das Evangelium nach Matthäus 4:435–36.

Indeed, this motif is also found in Matt 8:11–12, where, however, it is used very differently. The verbal parallels between Matt 28:16–20 and those biblical texts describing pilgrimage to Zion that could be relevant here are extremely slight. Finally, I would place in this category the possible allusions to Deut 31:23; Jos 1:1–9; 1 Chr 22:1–16; and Jer 1:4–10 pointed out by Dale C. Allison.47

Allison, The New Moses, 262–266; see also Davies and Allison, The Gospel according to St. Matthew, 3:677–87.

These texts display some structural analogies to Matt 28:16–20.48

It is therefore not advisable to speak about a biblical genre of “commissioning stories” (contra Benjamin J. Hubbard, The Matthean Redaction of a Primitive Apostolic Commissioning: An Exegesis of Matt 28:16–20 [SBLDS 19; Missoula: Society of Biblical Literature, 1984]).

The verbal identities, however, are limited to the verb

, “to command,” and to the motif of God being “with you”

. “All that [someone] commands/has commanded”

plus a form of

occurs thirty-six times in the LXX, and so it is impossible to speak about the influence of a specific biblical intertext here. Likewise, the Immanuel motif is widespread in the Hebrew Bible, and it is important to the whole of Matthew's Gospel.

49

See n. 33, above.

Once again, it is not possible to determine a specific biblical intertext to which the Gospel text was intended to allude. Rather, the Matthean text makes it possible for readers to recall every biblical text where the phrase

is predicated of God or where “all that I have commanded” occurs.

50

The “Immanuel” motif and

point to a christology of the presence of God and not to a christology of Jesus as a new Moses (contra Davies and Allison, The Gospel according to Matthew, 3:679–80).

The Matthean text gives the reader freedom to discover allusions to various biblical intertexts; such allusions were consequently “discovered” by biblical exegetes in the course of the history of interpretation, but misconstrued as intentional allusions. 4) The Septuagintal character of Matthew's language in general produces many biblical echoes, including formulas like

(“all the nations”), and

(“all the days”), as well as others. These echoes contribute to the biblical character of the Gospel and create for readers the impression that they are wandering through a biblical narrative landscape.

Again, on the narrative level, the close of the Gospel is comprehensible without any knowledge of its biblical depth-dimension. This depth-dimension, however, does exist, and it is very important. In these final verses, the motif of “God with us” is more important than the obvious and intentional allusion to Dan 7:13–14. Using the categories of Pfister, we may say that the latter has a high specificity, while the former has a very low specificity, because no specific biblical text is invoked. But the degree of communicativity and even of implicit dialogicity is much higher in the case of the Immanuel motif, because Matthew has used this biblical motif as a Leitmotiv throughout his Gospel. Furthermore, it is again true that this text, infused as it is with biblical language, enables and even encourages its readers and hearers to create their own biblical connotations, without directing them to specific intertexts. Biblical intertextualizations of Matthew's text by readers are among the creative acts of reading that the text makes possible.

It should be clear that both the intentional biblical allusions and the allusions whose “discovery” is enabled by the biblical character of the text are very important for an understanding of Matthew's Gospel. These allusions demonstrate that the Matthean story of Jesus is deeply rooted in biblical tradition; yet at the same time it is an entirely new story that records a new action of the biblical God. It has a biblical depth-dimension that cannot—and should not—be definitively grasped by its readers.

Such allusions also show an element of formal continuity between the Gospel narrative and the process of transmitting Israel's foundational history as it is recorded in the Bible and in early Jewish texts. There is a fundamental difference, however, between Matthew's unmarked absorption of Mark's story of Jesus and Matthew's unmarked biblical allusions: Mark's text is transmitted and renarrated in its entirety insofar as it serves as the hypotext of Matthew's Gospel. The unmarked allusions to biblical intertexts normally have no function as hypotexts. Those unmarked intertexts do not structure the Gospel of Matthew; rather, they appear—scattered and usually independent of each other—within a new foundational story. The adoption of that new foundational story is characteristic of the communities of Jesus followers, but it is not, of course, a feature of early, non-Christian, Jewish texts. Perhaps we may say that the biblical texts evoked by the intentional allusions are a secondary matrix for Matthew. His primary matrix, however, is the foundational story that he found in his new hypotext, the Gospel of Mark.

Quotations

I begin with a definition. A quotation—unlike a mere allusion—includes for me the conscious and “extensive, word-by-word” appropriation “of a longer given wording”;51

Koch, Die Schrift, 13.

nevertheless, the degree and the extent of verbal identity that distinguishes a quotation from an allusion cannot be defined mechanically. The difference between allusion and quotation is fluid. The absence of an introductory quotation formula should not be a factor in evaluating a putative quotation; many quotations, particularly in Hellenistic literature, are not introduced by such a formula.52

Ibid., 12.

In the Gospel of Matthew, however, most biblical quotations are formally introduced by a quotation formula. Matthew follows Jewish custom in this respect.

The vast majority of Matthew's biblical quotations are spoken by Jesus himself. This makes it clear for the reader that Jesus, “your only teacher” (Matt 23:8), is quoting and interpreting the Bible continuously and consciously. Most of the biblical quotations spoken by Jesus refer to the interpretation and the praxis of the Torah: Jesus “fulfills” the Torah and all its righteousness through his teaching and his life. Of the eight explicit quotations of the prophets and the Psalms in Matthew, six are predictions (Matt 11:10; 13:14–15; 21:16, 42; 22:44; 24:15), and two occur in a polemical context (15:8–9; 21:13). It is possible to contrast Matthew's demonstration of Jesus' fulfillment of the Torah with his presentation of Jesus' fulfillment of prophetic texts. The fulfillment of the Torah is continuously proclaimed by Jesus himself through his teaching and his life. The fulfillment of the prophets, however, is stated by Jesus only occasionally. The vast majority of the prophetical intertexts quoted in Matthew's Gospel are cited by the evangelist himself, especially in his fulfillment quotations.

Fulfillment quotations

A large number of quotations from the prophets have been specially tagged. They are introduced by the fulfillment formula, which was probably inspired by Mark 14:49, but specifically composed by Matthew. The evangelist inserted these tagged fulfillment quotations throughout his book, but he concentrated them in his prologue for didactic reasons. Thus could he make clear to his audience, from the very beginning of his work, that the whole story of Jesus is one continuous fulfillment of prophetic predictions.53

For my view of the fulfillment quotations, see the excursus in Luz, Matthew 1–7, 156–64.

All fulfillment quotations are comments made by the narrator.54

The only (relative!) exception is the prophetic quotation formulated by Jesus (Matt 13:14–15), where the introductory formula resembles the fulfillment formula.

Such an abundance of formulaic comments by a narrator is unique in Jewish narrative literature. According to the criteria of selectivity and communicability, the fulfillment quotations are among the most pointed biblical intertexts of the New Testament.

What triggered Matthew's invention of the fulfillment quotation formula? Two factors compelled the followers of Jesus to narrate his story as their new foundational story. The continued parting of the ways with Judaism was one factor; a second was the new perception, by those outside the movement, of Christianity as a distinct and novel religion. Christians were motivated to reflect on their own identity and to remember anew their own beginning, their

(Mark 1:1).

55

Such reflection and remembering is evident in the Gospel of Mark, a product of the Neronic persecution in Rome, when Christianity was proclaimed to be a new, specific “superstition” (Tacitus, Ann. 15.44.3) different from Judaism. Gerd Theissen (Die Religion der ersten Christen [Gütersloh: Gütersloher Verlagshaus 2000] 233) correctly interpreted the writing of the Gospels as a decisive step towards the final separation of the Jewish and Christian faiths. The early Christians started to write their own foundational story, and this marked their departure from the narrative community of Judaism.

In Matthew's Gospel, the painful experience of the parting of the ways elicits a bold response: The Matthean church programmatically claims the prophetic heritage of Israel as the legitimation for its own new foundational story.56

Luz, Matthew 1–7, 161–62.

Considering the title of Matthew's Gospel (1:1) and its other echoes of its “secondary hypotext,” the Bible, we can say that such nods toward the biblical text converge with the fulfillment quotations in their intention. By means of his title and the pentateuchal structuring elements of his book, Matthew aims to give his story of Jesus a Bible-like dignity. By means of the fulfillment quotations, he confirms that, in his story of Jesus, the prophetic promises of God have now been fulfilled. These two methods of appeal to biblical intertexts constitute the rhetorical strategy by which Matthew advances a claim to quasi-canonicity on behalf of his new foundational story. His portrayal of Jesus as one who quotes the Bible with authority strengthens this claim.

Concluding Remarks

1. In the Gospel of Matthew there are two entirely different types of intertexts. The first is the Gospel of Mark, the primary hypotext upon which Matthew's hypertext is based. It remains invisible to Matthew's readers. In spite of its invisibility, the Gospel of Mark completely determines the structure of the Gospel of Matthew and functions as its primary matrix. Matthew as a whole can be interpreted as a metatext of Mark in a rather strict sense of the word. This is not true for its relationship to Q, however, which merely supplements Matthew's “primary matrix.”

The case of the Bible as intertext is very different. It serves more generally as the basic text of reference that illuminates and interprets the new Matthean foundational story. Beyond that, it enhances the status of Matthew's Gospel by giving it a quasi-canonical character.

2. In my opinion, it is not possible to speak about a continuing process of biblical tradition that links the Bible and Matthew—in spite of the close analogies between the ways in which biblical narrators, early Jewish narrators, and Matthew himself retell their stories and employ biblical materials and patterns. Even though Matthew borrows basic structures from biblical literature in order to construct his own book as a “new Genesis,” such continuity is eclipsed by his adoption of a new foundational story. Consequently, the foundational history of Israel serves Matthew not as a constitutive “primary matrix,” but rather as a “reference text” subordinated to his new foundational story, the story of Jesus.

3. The very numerous allusions to the Bible throughout Matthew's Gospel indicate that the Bible functions for Matthew not only as its interpretational “reference text,” but also as its “secondary matrix” insofar as it lends to that Gospel its biblical character. Because it is permeated by innumerable biblical background-texts and suffused by countless biblical echoes—some of which were intentionally invoked by the author, some of which are only discovered by the reader who explores the biblical character of the whole Gospel—Matthew's story of Jesus acquires a biblical depth-dimension. Its readers may thus conclude that the God of the Bible is at work in the life of the Immanuel, Jesus, in a very intricate fashion.

Footnotes

This was first presented as a lecture at Harvard Divinity School, 1 December 2003.