Within the realm of biblical interpretation, canon is a term deployed as a key notion in how meaning is made, from ancient writings through the creation of a privileged set of authoritative texts. As such, canon belongs to a larger pattern of human meaning-making, through setting aside sets of objects, practices, or wordings as privileged. According to Jonathan Z. Smith, this identification of a privileged set exists in close coordination with the establishment of a cadre of persons who are called upon to apply the limited set to the infinite variety of human experiences. The tension between the unending field of that which is to be understood and the limited frame of objects, practices, wordings, or texts useful for meaning elicits ingenuity.

In this way, the basic structure of canon as a larger pattern of human meaning-making is, according to Smith, divination. That is, the practice of using a limited number of objects to divine meaning from an endless variety of occasions provides the basic frame for understanding how canon functions for a wide spectrum of religious systems. For instance, the southwestern African diviner with a basket of twenty-four objects uses just those objects in instruction for his clients as they negotiate a large variety of questions and dilemmas about their daily lives. For biblical interpretation, the limited number of interpretive means (i.e., the objects in the basket) are the privileged set of books. Exegetes, hermeneutes, and other castes of interpreters, then, are ingenious agents in the production of meaning from the limited set of texts with reference to an unending variety of human circumstances.

One can note that placing the biblical category in a larger frame of human patterns of behavior does not diminish the canon, but helps identify the particular creative activity inherent in the establishment, management, and practice of privileging a limited set of texts for an ever widening field of human experience.

This also coheres with the etymology of the Greek kanon and the Latin canon. The basic meaning in these ancient languages is that of a stick or rod used for measuring. It also carries the tangible meaning of charts or lists. By extension it can connote metaphorically standards by which something is judged or the judging itself. In some cases, this extends to the designation of a judging authority. So the ancient (that is, Hellenistic and Greco-Roman) references to kanon do not really indicate a collection or even a list of privileged books, but a much larger and more vague category of measurement, with the physical act of measurement dominating over the metaphorical.

For Smith the creativity and meaning provided in the dynamic of canon has almost nothing to do with the quality of the set of “objects” selected. Indeed, that which makes up the list is by and large “arbitrary,” although it may in some vague sense represent “preoccupations” of the groups making the selection. The creative ingenuity is in fact heightened by this arbitrary character of the set of, in the case of written canon, books. Nor is the privileged set of objects or texts “closed” or “static.” (Smith 1982, p. 44)

This helps make sense of the ways the biblical canons are in fact multiple, have much material that is sloppily repeated, are difficult to decipher in terms of the criteria of the selection of books, and have not been stable throughout history. The power of meaning created by canon does not depend on the privileged set of books remaining the same. Rather, meaning is made primarily through the endless exhibition of different meanings within limited frames.

The Notion of Canon in the Biblical Traditions.

Although the notion of “The Bible” itself—whether Hebrew, Protestant, Catholic, Syrian, or Ethiopian—derives from the concept (and/or practice) of canon, it is surprising how late these canons actually developed. The traditions of Israel took more than 750 years even to formulate the most rudimentary set of privileged sets of texts. Similarly, Christianity, with a much stronger built-in legacy of the prestige of that which is written, took at least 400 years for its provisional and regionally-based sets of privileged texts to develop, only to see them remain much less stable than those of Judaism.

David Carr has suggested that the notions of canon for both the traditions of Israel and early Christianity derive primarily from the understanding of Greco-Roman cultures that there was a kind of circumscribed set of classic (mostly) Greek and Roman texts that were required reading for those who were civilized. According to Carr, both initial collections and lists of Israel-related texts and early “Christian” texts need to be considered counter-collections to the classic Greek and Roman texts. This helps understand the relatively relaxed and occasionally loose nature of the protocanons of both spiritual Israel and proto-Christianity in that the notion of classic Greek and Roman texts was only loosely constructed and open to debate.

Inasmuch as Carr’s suggestion has merit, such collections within spiritual Israel and proto-Christianity would have been conceived at a somewhat more applied level of resistance than the larger grasp of canonizing tendencies of J. Z. Smith, which includes—but is not limited to—strategies of resistance. In addition, it is important not to perceive the collections/lists of proto-Christianity and spiritual Israel as mutually exclusive. Rather they need—at least from the second to fifth centuries—to be seen as overlapping and based within a larger strategy of countermand and resistance on the part of several kinds of spiritual Israel (and that would include proto-Christianity). The authoritative character of Septuagintal and Hebrew scriptures for a wide range of Christ movements of the first and second centuries would be the primary indication of such overlapping functions and identity.

The Development of the Canon of the Hebrew Bible.

There is no real indication of privileged collections—or even lists—of texts within the inter-tribal or monarchical periods of Israel. There were, of course, libraries and collections of texts: the somewhat legendary accounts of King Solomon in the tenth century B.C.E., King Josiah in the seventh century B.C.E., and Judas Maccabeus in the second century B.C.E. are probably relatively reliable, but little sense of selecting out some of them from others as having primary authority is evident. Even the first designations of privilege for Torah as a specific and guarded collection seem to have occurred in the postmonarchical era. This eventual designation of the authority of a given set of Torah texts, of course, belongs, then, to the more general rise of the importance of texts in the second Temple/colonialized period of Israel, from perhaps the late sixth century B.C.E. through the fifth century C.E. It is difficult to identify exactly when the first privileged collection of Torah texts took place, but the process within colonialized Israel of elaborating and redacting the Torah would not have allowed a designation of such a Torah collection, as it is now known, much before 300 B.C.E.

It is by and large agreed that the first major collection of such privileged texts from Israel actually took place in Egypt with the production of the Septuagint, a collection of the translation of Hebrew texts into Greek accomplished sometime between approximately 300 and 50 B.C.E. by many thinkers within the community of spiritual Israel living in Egypt at that time. The Septuagint contained 45 to 51 books (depending on how one counts the various books as they were included in this collection), the vast majority of which were eventually contained in the later collection of Hebrew scriptures.

The first-century historian Flavius Josephus noted that scriptures of Israel were divided into three types, but did not enumerate what books belonged to each of them. He did say that there were 22 books, 5 of which were the Law, 13 of which were prophetic, and 4 others containing hymns and rules of life.

It is only after the destruction of the second Temple in Jerusalem that rabbinic leaders began to focus on sets of privileged Hebrew texts. By this time, they were also inevitably under the influence of the existing Septuagint. Much of that influence, however, was negative, in that the Septuagint’s Greek language had come to represent the Greco-Roman tyranny over Israel. What followed were several hundred years of efforts to establish clearly a privileged collection in the Hebrew language. These efforts were done in the blossoming rabbinic movement, which eventually defined the traditions of Israel for the ensuing years until the present day. The rabbinic collections of Hebrew texts are in interesting tension with similar Hebrew texts collected by the communities at Qumran in the first century B.C.E., in that both the wording and the collections themselves differ from one another. Ben Sirach in the second century B.C.E. and Qumran probably in the first century B.C.E. both seem to have been aware of a privileged status of five books of Torah, Isaiah, Jeremiah, Ezekiel, and a group of twelve other prophets.

In Josephus, Qumran, and the rabbinic foundations, a criterion for giving some texts privilege over others seems to have been a shared notion that nothing after the first destruction of Israel in 587 B.C.E. could have been in such a (vague) privileged circle of texts. This is ironic in that, as noted earlier, the plurality of biblical scholars hold that even the Torah was not completed until well after this date. This most likely resulted in an effort to complete, redact, or compose literature in the colonialized, second-Temple period that had voices meant to approximate those of early Israel.

It is still very difficult to determine when the rabbinic movement completed the task of designating a definitive set of privileged texts one might retrospectively call “canon.” It has been held that it occurred at the “council of Jamnia (Yavneh)” in Galilee in the mid-80s C.E., but this opinion no longer holds sway in most scholarly circles. Rather it seems as if this process was brought to closure somewhere between 200 and 400 C.E. In the Mishnah, most likely from the second century C.E., Yadaim 3:5 tells of debates on the status of the Song of Songs and Qoheleth. Unfortunately the documents that to some extent chronicle such choices are not themselves easily dated. These occasional and partial observations present themselves in their current state in a larger collection of Mishnah and Talmud, which hold texts alongside each other without distinction from beginning in the second century B.C.E. through at least the fourth century C.E.

Whenever the rabbinic process was completed, it designated 24 texts as privileged “Tanakh” beyond others. The 24 were 5 books of Torah (Genesis, Exodus, Leviticus, Numbers, and Deuteronomy), 8 books of prophets (Joshua, Judges, Samuel, Kings, Isaiah, Jeremiah, Ezekiel- and the Twelve Prophets, which include Hosea, Joel, Amos, Obadiah, Jonah, Micah, Nahum, Habbakuk, Zephaniah, Haggai, Zechariah, Malachi), and 11 Writings (Psalms, Proverbs, Job, Song of Songs, Ruth, Lamentations, Qohelet, Esther, Daniel, Ezra–Nehemiah, Chronicles). The word Tanakh comes from the first Hebrew letters of each of the three parts of Torah, Prophets, and Writings. The book of II Esdras, dated from the late-first through the mid-second century C.E., is the first book to acknowledge a privileged collection of 24 writings, making clear that at least some rabbinic authorities worked with those 24 privileged texts by the second century C.E. On the other hand, the ongoing awareness in Yadaim of contestation on the status of Qoheleth and Song of Songs in the second century seems to indicate that in some quarters the circle of these 24 texts was still unclear.

Although to some this concludes the development of the privileged “canon” of the traditions of Israel and what came to be called “Judaism,” for most contemporary Judaism it is not so simple. Indeed, the privileged collections of Talmud in many ways have a kind of interpretive authority that surpasses that of the “Hebrew Bible.” The Talmud is usually considered to have two parts: the Mishnah (perhaps concluded around 200 C.E.) and the Gemara (perhaps concluded around 525 C.E.). The Mishnah reflected on earlier versions of the Torah, prophets, and writings, expanding them and providing limitations on how they might be interpreted. The Mishnah also felt relatively unfettered in adding dimensions of law and meaning, and is in most circles considered to be a part of what is known now as Jewish scriptures. The Gemara worked more to codify the books that eventually came to be the Hebrew Bible and the Mishnah. As one contemplates the various authorities and privileges of Hebrew Bible and Talmud, it becomes clear that the term “Bible” itself does not take seriously enough the complex circles of privileged literature in Judaism.

The Development of the “New Testament.”

The development of privileged collections of writings for the Christ movements is, at least for the first hundred years, more or less the same as for other traditions of Israel. From the earliest writings of Paul and the earliest gospels in the traditions of Mark and Thomas, there is evidence that the Septuagint was available and considered to have considerable authority as a collection. There seems to be little indication of any privileged collection of Hebrew scriptures in the life of the first century Christ movements, since their literature quotes only the Septuagint.

The collection now known as “The New Testament” did not exist even in an implicit form for at least 330 years after Jesus died. For those first three hundred plus years, the only hints of something that might eventually become the New Testament were various lists from different Christ movement writers and the regional “synods” of the movements. These lists of what early Christ followers should read did not agree with one another. As lists proliferated and Christ groups grew in numbers over the first several centuries, some groups started collecting the actual writings of Paul and others, but none of these collections corresponded to the 27 books that eventually would become the New Testament. In the fourth century, some church officials and groups began to try to propose what the best readings for early Christians were. But no historian—either conservative or liberal—can find even the idea of a collection of books called the New Testament, much less an actual collection of books, before 400 C.E. It was perhaps in the fifth century that the first collection of books actually corresponded to the 27 books of the New Testament. There is no evidence of an actual book called the New Testament or the (Christian) Bible until at least the eighth or ninth century.

The Lists of Recommended Books.

The first person to list what early “Christians” should read seems to have been the theologian Marcion. Around the middle of the second century, Marcion declared that everything written for Christ groups was in error except for the works of Paul and certain parts of the Gospel of Luke. The errors which concerned Marcion had to do with what he perceived as an overdependence on the traditions of Israel in books like the Gospel of Matthew, the Gospel of Mark, and the Gospel and letters of John. It seems relatively certain that Marcion did indeed make such a list, but no writings of Marcion survived the accusations of heresy against him. Nor did the opponents of Marcion produce lists of their own.

Perhaps the most assertive opponent of Marcion was Irenaeus, the bishop of Lyons, France, who wrote in the last several decades of the second century and the first decade of the third century. Irenaeus’s critique of Marcion had little to do with Marcion’s list of privileged books. Without making his own list of books, Irenaeus seems to be the first thinker to have used the term “The New Testament” as an inexact term for “Christian” documents in distinction from the Hebrew scriptures. Irenaeus also seems to have been aware that there were a number of Gospels. In response to either Marcion’s insistence on the one Gospel of Luke or the existence of a number of Gospels, Irenaeus was the first to designate Matthew, Mark, Luke, and John as needing to be considered privileged Gospels. Irenaeus’s central point in defense of these Gospels is that one must have four Gospels since there are four corners to the earth and the wind blows from four directions.

Another implicit endorsement of these four Gospels also occurred in the late second century. The Syrian theologian, Tatian, did not so much designate them as privileged as use them as sources for a cut-and-paste story of all four Gospels. This new amalgamation of Matthew, Mark, Luke, and John was probably in response to rising concern among the Christ people about differences among the many gospels. The other “list” of Gospels from the late second century was from Clement of Alexandria. He rejected the full authority for Mark and Luke, and accepted Matthew and John because only those two were at that time thought to be written by apostles.

In the period 190–310 C.E., there seems to have been little interest in establishing a privileged list of books. Nor did the Council of Nicea in the early fourth century produce any such list. However, Constantine’s commissioned historian, Eusebius, did produce a list as a part of his writings. Eusebius created three categories of early Christian literature:

  • 1. “recognized” (Matthew, Mark, Luke, John, Acts of the Apostles, 14 letters of Paul, I Peter, I John),
  • 2. “disputed” (James, Jude, II Peter, II John, III John) and,
  • 3. “spurious” (The Acts of Paul, the Apocalypse of Peter, the Teaching of the Twelve Apostles).

The Revelation to John seems to have been both “recognized” and “spurious” in Eusebius’s somewhat confusing discussion of it.

This uncertainty in Eusebius’s discussion of the Revelation to John highlights the character of Eusebius’s lists. Although some twenty-first century writers cite his lists as authoritative, in using the categories of “recognized, disputed, or spurious” it is not clear that Eusebius is making judgments about the value of the different books so much as reporting on how the books were treated.

As it unfolded, the fourth century was eventful, if inconsistent, in regard to the eventual New Testament. Many historians place the Muratorian Fragment in the fourth century, although it can be dated anywhere from 180 to 800 C.E. This document without a known author lists Matthew, Mark, Luke, John, the Acts of the Apostles, 13 letters of Paul, Jude, “two (letters) with the title John,” and the Wisdom of Solomon (a book without any mention of Jesus and almost certainly written before Jesus). It does not list either of the letters of Peter, III John, or the Epistle to the Hebrews.

The regional church council of Laodicea in Asia Minor met in 363 and 364. There are no written records of it, but later reports contain two different lists from the council, one recommending the same 27 books as are in today’s New Testament, the other recommending 26 books and leaving out the Revelation to John. There was another regional gathering in Hippo, North Africa, but here too there were no records of what was decided. A later council in Carthage reported in 397 that the Hippo “synod” had produced a list of the same 27 books known today as the New Testament.

For most scholars of New Testament canon, the decisive moment for the formation of the New Testament came in the spring of 367 C.E. when Athanasius, a bishop in North Africa, wrote a “festal letter” at Easter, instructing the Christians of North Africa to read 27 books as their authoritative literature. As in Hippo and possibly Laodicea, these 27 books were the same books as eventually came to be the New Testament known today. Athanasius calls this list the “The New Testament” and uses “canon” to refer to his list. This letter was not meant for all of Christianity at the time, but just the Christians of North Africa. Nor did Athanasius provide any of the churches with copies of these 27 books.

The Apostolic Constitutions, most likely written in the early fifth century, listed everything on Athanasius’s list except the Revelation to John, but added I Clement, II Clement, and 8 books of the Constitutions themselves. This final list discussed here demonstrates, then, some increasing agreement at the beginning of the fifth century on the authoritative status of the 4 Gospels, the Acts of the Apostles, 13 letters of Paul, Hebrews, the 2 letters of Peter, the 3 letters of John, and Jude; but it destabilizes Athanasius’s list of 27 with its own list of 36 books.

One complex issue around this discussion of ancient lists that point toward the eventual New Testament is its exclusion of a major nonwestern part of ancient Christianity. While western Christianity grew toward its Catholic and Protestant conceptions of the New Testament, the churches of Syriac and Ethiopian origin treated all of these issues quite differently. These churches developed different and more expansive lists, and even today differ with western Christianity as to what scriptures are authoritative. The Syriac version of the New Testament, known as the Peshitta, does not include II Peter, II and III John, Jude, and the Revelation to John. The Ethiopian New Testament includes books otherwise not a part of other New Testaments: The Sunodos, The Octateuch, The Book of the Covenant, and the Didascalia.

The Collections of Christian Literature.

During the first five centuries of Christ communities, there were a number of collections of early Christian literature. In contrast to the various lists, these collections had little relationship to the eventual New Testament. They do demonstrate that for some communities the actual collection of literature became important.

It is likely that by the end of the first century there were at least several collections of the letters attributed to Paul. Since the production of books themselves only began in earnest during the second and third centuries, most of these Pauline collections must have been in scroll form. Who made these collections of Pauline literature is less clear. They could have been from the more established Christ communities in their second and third generations. Or, they could have been collected by some very early Christian schools.

It is quite certain that in the second century a number of “Christian” schools emerged. Some of these schools had at their disposal various written works, both from the “Christian” writers of the first and second centuries and often from important philosophical writers as well. These school collections almost certainly contained works from the eventual New Testament and works outside it. By the fourth century, such school collections could have been relatively large. But there seems to have been little interest in these collections corresponding to the various lists of recommended instructions for “Christians” at large. And, by the fourth century a considerable number of “Christian” communities seem to have collections of literature to read in worship and study settings.

In possibly the third century and certainly the fourth century, monasteries also became centers of considerable study, and they began to have libraries within their building complexes. One of the largest and earliest collections of this sort is the Nag Hammadi “library” found in 1945 in a jar several miles from a fourth-century Pachomian monastery in central Egypt. One of the most interesting aspects of the Nag Hammadi collection is that it is in the form of actual books. From other non-Christian libraries of that time, it seems clear that by the fourth century such a monastic library would have probably contained both books and scrolls.

Relative to the New Testament it is probable that some of these collections included all 27 books of the eventual New Testament. On the other hand, it seems unlikely that any of these collections contained only these 27 or even grouped them together as a “New Testament.” Indeed, it seems improbable that there were any separate collections of the 27 books of the New Testament in the first 400 years of Christian tradition.

The Actual Production of New Testaments.

Although we have very few actual examples, it is almost certain that some of the collections of “Christian” literature of the fourth century had multiple little “books” in actual leather bound form. For instance, the collections found at Nag Hammadi consist of leather bound books, each of which contains 3–10 separate works.

There are two reasons to doubt that the first 500 years of “Christianity” produced one book called the New Testament that included the specific 27 little “books” that now make up the New Testament. First, the technology of book production was such that putting all 27 “books” into one was most probably impossible. Second, as seen in the above study of the lists of books for “Christians” to read at that time, there was not consensus on what the “mandatory” New Testament content was until at least the fifth century.

Around the year 405, it seems likely that the theologian and translator Jerome had a unified translation of the 27 documents of the eventual New Testament. Later versions of the Vulgate contain an introduction to the translation most likely written by Jerome himself. This introduction both assumes and declares that the Vulgate translation of the 27, and only 27, books has special authority for Latin Christians. So although it was not actually one book, both its unified translation and its introduction come quite close to it being an actual book called the New Testament.

The first New Testaments—as an actual book—were probably produced within two or three centuries of Jerome, as the Jerome’s translation of the 27 gained authority in the churches. This production of an actual book of the New Testament somewhere between the seventh and ninth centuries signaled on one level the end of an at least 600 year process of the making of the New Testament as we know it. On another level, this consensus had a number of limits. Perhaps most important is that this “New Testament” was only for those who read Latin. In the seventh through ninth centuries of flourishing Syrian and Ethiopian Christianity the scrolls and codices did not follow the consensus of Latin-speaking North Africa and Western Europe. Rather, in these regions authoritative collections of Christian books were not the same. Even today, the fifth-century decisions of a major part of Syrian Christianity to have a New Testament contain just 22 books (excluding II Peter, II John, III John, Jude, Revelation of John) are still held to be authoritative; and for many centuries Tatian’s Diatessaron (described earlier) replaced the four Gospels.

In the ensuing 1,000 or so years (450–1500 C.E.) in both the east and west, many different versions of the New Testament were produced. The vast majority of these New Testaments were in Greek and Latin, languages that no one spoke in most of the areas where Christianity existed.

With the advent of the printing press, the New Testament became available to the broader population beyond those who read it in Latin and Greek. With translations into the common tongues and the mass distribution via the printing press, rather suddenly the New Testament was of direct interest to common folk.

This change in public consciousness produced several responses from church leaders. Martin Luther tried unsuccessfully to have four books—Hebrews, James, Jude, and Revelation—removed from the New Testament as read by the new Protestant public. And, in the sixteenth century, several church-wide bodies (including the Roman Catholic Council of Trent) for the first time declared that 27 specific books officially belonged in the New Testament.

Canonical Criticism.

Since the mid-1970s some biblical scholars have participated in what has come to be known as “canonical criticism.” The term itself has not been able to focus energy or a collegium of scholars. Nor have people been able to agree on either approaches or conclusions. Nevertheless, it bears noting that even in their disagreement a number of scholars have brought the phenomenon of textual canon to the attention of broader spheres of interest. Together these efforts and disagreements can be characterized as pointing to the way canon as a textual phenomenon convenes particular kinds of meaning. At points this interest in canon in itself as a particular kind of meaning has been as broad as simply acknowledging that closed sets of texts make particular kinds of meaning, while at other junctures this interest in canon has been in the service of an assertion of a particular kind of Christian meaning. Still others have been interested in the ways certain sets of texts have influenced either the course of history or the course of biblical scholarship. More technical aspects of this loose and vague field of canonical criticism have concentrated on how the biblical canon(s) have been formed.

The name most often associated with canonical criticism has been Brevard Childs, although he himself rejected the term. Childs’s work has shown great scope, and dealt with technical, philosophical, theological, and—one might say—even apologetic dimensions of canon. The earlier work of James Barr is often seen as a predecessor to that of Childs. James Sanders’s more recent scholarship powerfully convened formal and technical interest in what canon meant as both an ancient and modern meaning-maker. In terms not dissimilar, Jacob Neusner and Burton Mack both deconstructed notions of canon formation and reflected on what was at stake in larger cultural investments in canon. Perhaps the most evenhanded and overarching of this work has been done by Lee MacDonald.

Although even the term “canonical criticism” mostly occasions disagreement, this clumsy and charged set of efforts can be credited with two achievements: (1) the unfocused identification of questions about whether “canon” is a naturalized category of meaning within the field of biblical interpretation; and (2) an almost unconscious match between these various assertions of meaning in and for biblical canon with Jonathan Z. Smith's keen observation that the creation of meaning is at the heart of any arbitrary human selection of privileged objects, ideas, or texts.



  • Brakke, David. “Canon Formation and Social Conflict in Fourth Century Egypt: Athanasius of Alexandria’s Thirty-Ninth Festal Letter.” Harvard Theological Review 87.4 (1994): 395–419.
  • Childs, Brevard S. The New Testament as Canon: An Introduction. Harrisburg, Pa.: Trinity Press International, 1994.
  • Gunn, David. “What Does the Bible Say?: A Question of Text and Canon.” In Reading Bibles, Reading Bodies: Identity and the Book, edited by Timothy K. Beal and David M. Gunn, 242–261. New York: Routledge, 1997.
  • Keefer, Kyle. “A Postscript to the Book: Authenticating the Pseudepigrapha In Reading Bibles, Writing Bodies: Identity and The Book, edited by Timothy K. Beal and David M. Gunn, 232–241. London and New York: Routledge, 1997.
  • McDonald, Lee M. The Formation of the Christian Biblical Canon. Peabody, Mass.: Hendrickson, 1995.
  • Smith, Jonathan Z. “Canons, Catalogues, and Classics.” In Canonization and Decanonization, edited by A. van der Kooij and K. van der Toorn, 295–311. Leiden: Brill, 1998.
  • Smith, Jonathan Z. “Sacred Persistence: Toward a Redescription of Canon.” In Imagining Religion: From Babylon to Jonestown, edited by Jonathan Z. Smith, 36–52. Chicago: University of Chicago Press, 1982.
  • Wimbush, Vincent L. “Reading Darkness, Reading Scriptures.” In African Americans and the Bible: Sacred Texts and Social Textures, edited by Vincent L. Wimbush and Rosamond C. Rodman, 1–22. New York: Continuum, 2000.

Hal Taussig