nthologies will always be with us.[1] But whatever the syllabi of foundational literary survey courses may tacitly proclaim, an anthology is not the same thing as "the canon." Like anthologies, the canon has inevitably to do with pedagogy (Levin 1981). But unlike anthologies, the canon is not necessarily transient and quickly dated. Nor is the literary canon able to be historically triangulated, I would argue, from all the literary anthologies that are in print at a given moment—a finer point to which I will return. The anthology is a literary storage and communication form: a textbook, (now) a digital archive, (once) a commonplace book, (perhaps still) the poems one has memorized for pleasure. The canon, on the other hand, is not a form, but a literary-disciplinary dynamic: it is a field of force that is never exclusively realized by any physical form, just as metal filings align with but do not constitute a magnetic field. Grasping this dynamic entails a process of concept-formation that both clarifies and complicates our current understanding of such basic but slippery concepts as "canon." Thus far, the formal definitions of the term "canon" number six (Fowler 1979) or even ten (Harris 1991): this multiplicity could use some conceptual simplification. Yet "the canon" is also something that all literary scholars and students tend to live with tacitly, fish in water: for conscious beings, such a life-supporting medium should appear at times opaque, rather than continuously transparent. What I offer here at the fork of conceptual simplicity and complexity is the addition of a third term, "corpus," to designate the other focus that works in summation with "anthology" in the field of "canon." I should emphasize at the start that I am not offering a structuralist binary opposition. What I am offering is the formation of a dynamic concept of the canon, that is, as a dynamically changing field of force, whereas structuralist concepts remain distinctly static. Now, the invocation of the scientific field of force as a way to form literary concepts is not new (Hayles 1984),[2] but I mean it here specifically in a sense perhaps less well known, that of the topological and field psychology of Kurt Lewin. I will outline Lewin's theory of scientific concept-formation here as the high road to forming the concept of canon as an anthology/corpus dynamic.

Lewinian Dynamic Concepts

It may not be going too far to say that Kurt Lewin has had a greater impact on contemporary psychology than both Freud and Jung. Though less remembered today than these contemporaries, Lewin (1890-1947) is a key founder of social and organizational psychology, and he originated such influential ideas as group dynamics and "cognitive dissonance."[3] Like the latter term, and like many of Freud's concepts, the keywords of Lewin's thought and experimental research have penetrated thoroughly into the contemporary mind. Some of these Lewinian ideas and their applications include "sensitivity training," leadership and parenting styles (autocratic; democratic; laissez-faire), "groupthink" and group decision, training groups ("T—groups"), "quality circles" and "total quality management," and the "social climate" and psychological "tension." Lewin's name is rarely associated with such concepts in large part because Lewin himself as an academic colleague was democratic to a fault, and allowed his students and associates to share his ideas and receive credit for them.

It may also be true that much of Lewin's theory is too difficult for general consumption—Lewin is certainly not half the writer that Freud is. In developing his influential but demanding field theory of psychology (Mey 1972), Lewin employed geometrical topology, the mathematics of areas (Leeper 1943). The result was an abstruse psychological system: with his complicated diagrams of vectors and Venn-like ovals, Lewin mapped the dynamics of given psychological situations as sets of interposed forces. Above all, Lewin emphasized the need for dynamic concept-formation in psychological research, and this dynamical emphasis of his theory, rather than its topological specifics—which for impenetrability recall Lacan—has contributed most to Lewin's enduring legacy in social psychology today.

At the risk of slighting the breadth of Lewin's experimental and theoretical work in psychology, I will choose to focus here on the pivot of Lewinian concept-formation. Lewin enunciates his scientific philosophy of the formation of dynamical concepts in a dense essay entitled "The Conflict between Aristotelian and Galileian Modes of Thought in Contemporary Psychology" (1935).[4] The piece, which Lewin wrote in order to introduce his thought to his American colleagues, is an underrated landmark in the philosophy of science: despite its title, the essay explores analogies of the two modes in physics and biology as well as psychology, offering a concise yet sweeping statement of the nature of modern scientific method since Galileo. "Modes of Thought" does however eschew any historical argument, though the Galiliean mode is generally taken by Lewin to be an advance over the Aristotelian (1-2, 4). Rather, the central argument is that both modes of scientific thought are still in active use, and that the concepts that each mode produces are not reconcilable.[5]

Aristotelian scientific concepts, says Lewin, are characterized by evaluative categorizations that are abstract, binary, and based on the regularity and frequency of given events, whose potential randomness is ironed out in the establishing of their essential nature (2-9, 13-21). Galileian scientific concepts, on the other hand, emphasize concrete functions over abstracted classes, so that there are homogenous scales instead of valuative dichotomies, and so that single events are not considered to be ordinarily fortuitous or irregular (9-13, 21-26). Both the Aristotelian and Galileian modes are "empirical," but the Galileian mode paradoxically connects the integrity of the individual case with the universality of the scientific law,[6] which remain strictly antithetical in the Aristotelian mode. For psychology, the benefit of such integration is clear:

For Lewin, one of the most important distinctions between the two modes is the way in which each relates "lawfulness" to the individual case. The Aristotelian mode refers to many "historical-geographical" instances, but blurs their individual peculiarities in the creation of the law that governs them. The Galileian mode insists that laws must be exceptionless, if they are to be "laws," and thus the most unusual events may hold the greatest interest for law-formation—including cases that never really occur at all, as in the "thought experiments" of physicists:

For Galileian laws, "[h]istorical rarity is no disproof, historical regularity no proof of lawfulness" (26). For Aristotelian laws, "it is always evident not only that physical and normative concepts are still undifferentiated, but that the formulation of problems and the concepts that we would today distinguish, on the one hand, as historic and, on the other, as nonhistoric or systematic are inextricably interwoven" (8). The Galileian mode makes an initial and strict separation between system and history.

This contrast between the Aristotelian and Galileian treatments of law and event sets the stage in "Modes of Thought" for Lewin's discussion of the concept of dynamics in the two modes (27-42). The dynamical concept of an environmental "field of force" is only available in the Galileian mode, according to Lewin, since "[f]or Aristotelian concepts, the environment plays a part only in so far as it may give rise to Ôdisturbances'" (28). The working of a force such as gravity, for example, demands an "environmental" understanding: "not only is the upward tendency of a lighter body derived from the relation of this body to the environment, but the weight itself of the body depends on such a relation" (29). According to Aristotle's entelechy, on the other hand, an object already holds within itself the seeds of what it will become, and only other objects can hinder the reaching of this goal—this is the extent of Aristotelian dynamics. In the Galileian mode, however, the setting begins to determine the interaction of forces on the object: "[t]he idea of investigating free falling, which is too rapid for satisfactory observation, by resorting to the slower movement upon an inclined plane presupposes that the dynamics of the event is no longer related to the isolated object as such, but is seen to be dependent upon the whole situation in which the event occurs" (29). In the psychological realm, Galileian dynamics means that the unfolding of an event must be monitored in all its particulars, moment by moment, using a "process differential," like a calculus for the plotting of curves (32-35). In other words, the Galileian mode prizes the entire situation, considered as a dynamically shifting set of mutually interdependent facts, and thereby requires "getting rid of the [Aristotelian] historical bent"—so important is the presently unfolding interdependence of the "facts," the currently active forces in the field (28-32). Herein is the paradoxical power of the Galileian mode, as Lewin puts it in his ponderous way: "with such functional concepts it is possible to go from the particular to the general without losing the particular in the general and thereby making impossible the return from the general to the particular" (35).

In summary, Lewin's preferred Galileian mode of thought "depends upon keeping in mind that general validity of the law and the concreteness of the individual case are not antitheses, and that reference to the totality of the concrete whole situation must take the place of reference to the largest possible historical collection of frequent repetitions" (42). The Lewinian challenge for a dynamic concept of the canon in literature is thus twofold: is there a generally valid "law" of the canon (as "canon" itself suggests) that can encompass all the individual cases where it appears? And, can this concrete yet general "canon law" be formulated not as a historically-conditioned rule, but as valid for particular cases of the canon? Both challenges are answered affirmatively within one metaphor for the canon that already contains the same dynamic concept of concretely situated forces: the glacier.

Glacial Canon, Dialogic Dynamics

Certainly, the canon (like any disciplinary dynamic) is hardly a swift-moving force, and this is why the "glacial" metaphor seems so appropriate for defining the canon: "the glacially changing core is a kind of diachronic canon (canon-9), to be distinguished from a rapidly changing periphery that could be called the nonce canon (canon-10), only a minuscule part of which will eventually become part of the diachronic canon" (Harris 1991, 113). These are the penultimate and ultimate of Harris's ten definitions, building on Fowler's six (1979), but they could both be folded into the glacier metaphor: atop a glacier's core are layers of new snow and "old snow," or "firn," which are year-old granules on their way to becoming solid ice; in glaciology, the meeting point of the zones of new accumulation and "ablation" (loss by melting, etc.) is called the "firn line." The other canon subdefinitions of Harris and Fowler could also be subsumed into the glaciological scheme: the pedagogical canon is like the useful stream of glacial meltwater; the critical canon involves glacial fieldwork; the popular canon is like glacial tourism. Like a glacier in the natural world, the canon has immense power to channelize a cultural environment; can pick up the most likely or unlikely things and carry them great distances ("erratics"); and exists on a superhuman scale, tending to dwarf us in size and to transcend our lifespans in its gradual advances and recessions. But there are also times when the movement of a glaciers can accelerate—when a glacier decides to "surge." These "galloping glaciers," as they have sometimes been called, suddenly shift from no or slow movement (on the order of a few yards per year) to a cataclysmic advance of miles in just weeks or months. Interestingly, glaciers that surge seem to do so in a regular cycle of long build-up and sudden release (van der Veen 1999, 323-24). The literary canon seems to be one of these periodic "gallopers." For instance, the canon has recently opened, quite remarkably, to recognize a fourth supergenre in addition to poetry, drama, and fictional narrative—namely nonfiction prose, or the essay (Bloom). To judge from the competing anthologies of Abrams and Greenblatt on the one hand (2000) and Damrosch on the other (1999), the anthological acceptance of the "essay canon" has indeed happened at a gallop—there are suddenly thousands of pages of nonfiction prose in the benchmark Norton and the Longman British literature anthologies.

By taking the glacial metaphor to such extremes, I am after an understanding of the canon in the way it works dynamically, in Lewin's Galileian mode, rather than epigenetically. That is, we tend to worry about glaciation in terms of its epiphenomena, the Aristotelian "good and bad" dichotomies—the rising and falling of the oceans, the historical record of climactic variation layered within the ice, the intrusion of icebergs in shipping lanes, versus the good of tapping such icebergs for potable water. For the canon, important epiphenomena include the ways its forms can oppress and exclude, or be liberating. But the way that a glacier or canon works internally is something else again, and this dynamic can only be comprehended as the locus of continually related forces, in the Lewinian and Galileian mode.

For the glacier, being on earth and subject to gravity means more than one thing, forcewise: on the one hand, there is internal pressure and downward compaction, leading to a spreading that glaciologists call "creep," and on the other hand there is the partial resistance of the surface, over which the glacier also experiences "slide." For the canon, being on earth and subject to the accumulating weight of the snow of literary production means experiencing two forces, which I am calling "anthology" and "corpus." With Galileian dynamics, the actions of both glacier and canon are comprehended as a summation of the forces working upon them at given moments. A recent comprehensive work on glacier dynamics by C. J. van der Veen is quite Lewinian in this sense (1999). In order to establish a dynamical "flow law" for glaciers, van der Veen employs a "force budget technique," where the forces acting on a glacier must total out to exactly zero (37-43).[7] The approach is very Galileian: van der Veen has a universal expression of glacial flow that depends on no particular record of glacial activity for proof, but which can nonetheless be applied closely to any existing glacier or ice sheet (309-436).

Van der Veen finds that the many glacial stresses seem to be divisible into two kinds: "a lithostatic component associated with the action of gravity, and resistive stresses that oppose this action" (ix). In the same way, out of the many individual forces that can act on canons, two kinds appear again and again: the corpus, which is perhaps the more gravitational, and the anthology, which is perhaps the more resistive force. What we typically mean by "corpus" is "a (full) body of work," not necessarily authorial, and "anthology" typically means "a (limited) collection." Anthologies and corpuses are already well-known as "things," so I will differentiate them adjectivally rather than as nouns:

Each of these canon-forces is like, but not fully realized by, the bibliographical forms that we associate with them. The experiences that each kind gives us are also typical of the canon-forces. Anthologies are easier, teachable, and fun, since they are attractively produced as "clear texts." Corpuses are far heavier, multi-volume tomes, filled with scholarly apparatus, and only specialists read them completely.

The relationship is not binary-oppositional, for there is a different kind of relationship here:

Anthologies are deponent in relationship to corpuses, "taken from (the grammatical counterpart)," as in the "deponent verbs" familiar to students of Latin: a deponent verb has a passive form with an active meaning, but with no corresponding active form to supply a passive meaning, which is thereby impossible for that verb. In short, anthologies are always taken from and indeed purport to look exactly like corpuses, but are in fact their force-reverse. But only corpuses are fully "ponent" or thetic, in the sense of an active whole with all the possibilities available. To bring this back to the Lewinian terms,

and so on. The Aristotelian "good and bad" dichotomies exist only insofar as they are anthological; it is a conceptual mistake, as I would have it, to fight over entire canons as over "good and bad"—which is like calling gravity good or bad.

It seems here that anthology and corpus could also be likened, respectively, to the glacial zones of ablation (wasting) and accumulation: what is "good and bad" about glaciers is not the glacier, or the snow and firn at the top, but what happens at the bottom: all that anthological calving, galloping, melting, and tilling. But this distinction threatens to become another Aristotelian concept. As their deponent relationship suggests, it is not helpful to consider anthology and corpus as "binary oppositions"—they are not at all exclusive forces, just as gravity is not exclusively (by nature) "opposed" to the intermolecular properties of solid water at high pressure. A close analogue in literary theory to the Lewinian and Galileian mode of thought is Bakhtinian dialogism. In Joseph Natoli's words, this sort of functional relations is "dialogic—grounded in the openings of the theory body—and not structural—grounded in the regular surface of binary opposition" (1987, xii). I will examine some of these dialogic openings in the body of the canon below. In dialogic terms, "anthology" and "corpus" designate two particular voices whose dialogue of force is "the canon." According to Tzvetan Todorov, better known as a structuralist apologist, in a dialogical situation "[o]ne does not abandon the idea of truth, but one changes its status or its function by making it a regulatory principle of exchange with others, rather than the content of one's own program" (70). In the three cases of canon-formation which I discuss below, the canon is also heard as "the encounter of two voices" (Todorov 1984, 71), namely the corporeal author's and the anthology editor's.

Case One: Polyclitus's Canon

"Canon" holds within itself at least one defining contradiction, perhaps best captured by the English word "standard": does it mean "evaluative measure" (Gk. kanon "rod, ruler, measuring stick"), or does it mean "a comprehensive list (of authors or of works)" (as in the "old standards")? The latter meaning is much later historically, but it has the power of deriving from the formation of the biblical canon (Barr 1983). Nevertheless, this modern meaning of "canon" as "authoritative list" does not appear before David Ruhnken's use of it in 1768 (Pfeiffer 1968, 205-207). On the other hand, as Pfeiffer also notes, the concept of a canonical "full list" is already there long before, and not just in the scriptural canon. There were the lists of accepted auctores for study in the medieval grammar course, and similar sets for every other area of medieval study for that matter (Minnis 1988). Long before that there were the (variable) numbered lists (pinakes, "indices") formulated by the Alexandrian scholars, notably the Nine Lyric Poets and the Ten Attic Orators.[8]

But there is also another, even older classical precedent for "canon" as an aesthetic formulation. In a lost treatise entitled The Canon, the bronze sculptor Polyclitus of Argos (fl. 460-10 B.C.) listed the mathematical principles of statuary proportion, exemplified by his own best work, which also became called The Canon: the statue often called today the Doryphorus (The Spearbearer).[9] This Canon also no longer exists, but is represented by many Roman copies of a figure of a young man which has a distinctly chiastic arrangement: as the figure steps forward, the right leg and left arm bear weight, and the left leg and the right arm are relaxed.[10] Here again, the canon is a self-contained field of interposed force. Moreover, this force field is also exerted between the statue and the treatise: the statue Canon is the self-contained and self-sufficient example, or the corpus; the treatise Canon is the disembodied listing that points to the corpus, or the anthology.

The case of The Canon is glacially dynamic in two ways. First, Polyclitus's statue, like others in his own corpus (and like the legendary ones of Daedalus), appears to be "moving," just about to step off the pedestal and into the observer's space. This is the glacier's material paradox, similar to that of common glass, of liquid solidity, and amplified here by the material in question, the bronze. Second, The Canon was immediately recognized as such in the contemporary artistic world: as the written sources attest, the Doryphorus "swiftly entered the consciousness of both the literary elite and its wider audience, becoming a handy simile, even a cliché, to be deployed as appropriate" (Stewart 1990, 1:266).[11] This apparent exception to the glacial mode of the canon suggests that there was a felt need for the dynamic, dual concept of canonicity that Polyclitus had doubly embodied. And of course The Canon was absorbed into the sculptural milieu as well, to judge from such take-offs as a Discobolus (Discus-Thrower) by the sculptor Naukydes, a contemporary of Polyclitus (Ridgway 1981, 242-43). Thus, in Pliny's words, the Doryphorus "was called The Canon by sculptors, who derive from it the lineaments of art as from a law [lege], and so [Polyclitus] is judged to be the only one to have ever created art itself in a work of art."[12] Just as the scientific law and the individual case meet in Lewin's Galileian mode of thought, so the aesthetic rule and the given statue meet in The Canon. A similarly formed Galileian concept occurs in the case of the poet Meleagar's contribution of a corpus to the Greek Anthology.

Case Two: Meleagar's Garland

My argument is that all anecdotes of the canon work like Polyclitus's dual Canon, a zero-sum force-budget of corpus and anthology. On this view, the biblical canon is, on the one hand, a two-volume anthology, a two-set selection that excludes other apocryphal possibilities, but on the other hand a complete and authoritative collection, all written by a single Author, whereby it can be considered the exemplary, thematic source of all doctrine. The forces of anthology and corpus are never singly epitomized in bibliographic forms; every glacier experiences both the inevitable draw of gravity and the changing resistance of the lithic bed. Likewise, there are no exceptions to the canon dynamic: every canonical exemplar is a case of budgeting the forces of anthology and of corpus. One of the best anecdotes of this glacial-canonical dynamic, actually a protracted series of such cases, is contained in the long history of the Greek Anthology, a collection of epigrams culled from over a millennium and a half of writing. In the words of Alan Cameron, the Greek Anthology as we now know it, containing more than four thousand individual poems in sixteen books, has become more an inclusive corpus than a selective anthology: "The Greek Anthology is one of the great books of European literature, a garden containing the flowers and weeds of fifteen hundred years of Greek poetry, from the most humdrum doggerel to the purest poetry" (1996, 102).

The historical peregrinations of the Greek Anthology have been definitively mapped by Cameron (1993). The story begins with the Garland of Meleagar (ca. 100 B.C.), a lost collection of about eight hundred Greek epigrams, a long-lived and flexible form of poetry. By the first century B.C., epigrams had already evolved for more than six hundred years, from their Archaic Greek beginnings as straight-forward inscriptions placed on tombs, statues, and even daily objects, to more bookish uses in Hellenistic times as occasional poems, especially ones erotic, epideictic (rhetorical wit), and sympotic (for the banquet—both roasts and toasts). All of these subgenres were represented in Meleagar's Garland. Around A.D. 40, another Garland was assembled by Philip of Thessalonica, with more up-to-date selections, including Romans who wrote in Greek. Such collections continued to be made long into the Byzantine era, including the secularly classicizing Cycle of Agathias (ca. A.D. 536), but Christian moralizing was now eminently possible as well.

At the beginning of the tenth century, all of the foregoing collections and kinds (and more) were omnivorously recombined in a gigantic anthology by Constantine Cephalas, a Byzantine schoolteacher. Forty years later (ca. A.D. 940), Cephalas's work was reedited in a codex that found its way to the Elector Palatine's library in Heidelberg, and this source, the earliest source for almost everything contained in all of the preceding collections, is known as the Palatine Anthology. (Modern versions of the Greek Anthology are numbered according to the Palatine Anthology's arrangement of fifteen books.) [13] The final major twist to this canon story came in 1301 with Maximus Planudes, a scholar who revised Cephalas's work once again, using sources other than the Palatine manuscript. This abridged version, called the Planudean Anthology, was the one popularly known to the later Middle Ages and to the Renaissance, when it was printed for the first time in 1494. (The Palatine codex was finally rediscovered by Salmasius in 1606.) Planudes' work remains significant because his version preserves some of Cephalas's material which does not appear in the Palatine manuscript (380 epigrams, now appended as book 16 of the Palatine Anthology).

As is so often the case with works from the classical era, we depend for the text solely on the vagaries of medieval canonicity, and thus for Cephalas it is "Christians first": the earlier pagans do not appear until book 4. This "lateness" also appears to be true, at first, for the Greek Anthology's title. The Greek word anthologion as a book title does not appear until the early second-century A.D. collection of Greek epigrams by Diogenianus of Heraclea, and the term was not associated with the full Greek collection until Cephalas. Diogenianus's collection also does not figure prominently in the history of the Greek Anthology. Nevertheless, like "canon" in Polyclitus, two millennia before the nominal appearance in 1768 in Ruhnken, the concept of "anthology" had been fully formed long before Diogenianus applied the term. This metaphor of a poetry collection as a gathering of flowers (Lat. florilegium; cf. Eng. "po[e]sy") was in place at the Greek Anthology's beginning, the Garland of Meleagar of 100 B.C.. Though Meleagar's Garland (Gk. Stephanos, "wreath, floral crown") is known solely through its medieval glaciation, the first and final poems of the Garland have been preserved, and they show that "anthology" is indeed the avowed aim of this collection of epigrams. Meleagar's proem (now Greek Anthology 4.1) lists several dozen of the poets that are included in the Garland, and to each is assigned a flower (or other plant). In what is commonly taken as the Garland's final epigram (now 12.257), Meleagar announces that he has finished his poetry book and, invoking again the titular metaphor, "with flowers has plaited a Muses' Garland of verse" (12.257.6). Meleagar pursues the catechrestic possibilities of the flower metaphor throughout the Garland: the flowers could be the chosen poets, the selected poems, or the loves who are addressed in the poems (e.g., "boy-flowers," 12.256). Apart from the thoroughness of his application of it, Meleagar is hardly being innovative with the floral image schema: all its possibilities, including the bouquet, were available in the Greek epigram and lyric, for tropes poetic, sexual, or otherwise (the metaphoric extensions including fruit, bees, honey, seasons, and so on). In the groundbreaking recent study Poetic Garlands, Kathryn Gutzwiller furthermore argues that Meleagar is far from original in creating the Garland, given the extensive evidence of deliberately designed collections that preceded him (1998, 33-35). Gutzwiller also points out that "anthology" is attributable long before to Nossis (fl. ca. 300 B.C.): "it has apparently gone unnoticed that by identifying her epigrams with roses in an introductory poem [now Greek Anthology 5.170] Nossis makes the first known metaphorical association of a collection of poetry with flowers" (79).

While Meleagar's Garland is only available in pieces through many intermediaries, it has been reconstructed in detail by Gutzwiller (276-322). (The assumption is that many later editors, like Cephalas, seem to have simply copied large "runs" of poems straight from earlier anthologies, a practice which is obvious from these epigrams' ascriptions to authors who are associated as a datable group with particular earlier collections.) In Gutzwiller's analysis, Meleagar's Garland was divided into four books, each with a different topic (erotic, dedicatory, funerary, and rhetorical books; see Gutzwiller 1998, tables 2-5). Within the four books, there were linked introductory, topical, transitional, and concluding sections, "in each of which the practice of editing becomes an aesthetic endeavor rivaling poetic composition itself" (Gutzwiller 1998, 279). Thus Meleagar has created an anthology that is a series of corpuses within corpuses. Moreover, the poet that Meleagar has chosen to anthologize the most is himself—about one sixth of what remains of the Garland are Meleagar's own poems—and Philip does the same in his own Garland, so that both anthologies are also the de facto corpuses of these poets. Thus the canon of the Greek Anthology is characterized by the canonizing of both the corpuscular poet and the anthological editor.

Alan Cameron's comment on the distinctiveness of Meleagar's Garland, characterized as purely "useless" and ephemeral, precisely distinguishes the poetic motivation of the corpus from the editorial motivation of anthologizing:

As Gutzwiller has shown, Cameron's argument here is extreme (1998, 33-34). But Cameron has nonetheless pinned down a key difference between the anthology, as perishable posy, and the corpus, which has the didactic possibilities of any scriptural accumulation. Meleagar's Garland was obviously far from perishable, for from its creator's perspective it had the very useful function indeed of preserving his own poetry! A perishable bouquet is also meant to be portable and presentable with one hand—and with Meleagar's Garland, in its original form as four separate papyrus book rolls, the Greek Anthology is already becoming a perennial flower bed.

Case Three: The Classic Hundred

There are no such things as unmixed anthologies and corpuses, in the sense of the "pure forces" which I am discussing. Like Meleagar's Garland and Polyclitus's Canon, the bibliographical realization of canons always reifies the working of both anthological and corpuscular forces. Polyclitus and his Canon are found in the corpuses that art historians have collected—all the available evidence, both the statues and the texts that comment on them, but these also function much like anthologies, since they hardly include every extant shard of statuary (Boardman 1985; Pollitt 1990; Stewart 1990). There is really nothing left of the Hellenistic epigram apart from Meleagar, so his Garland has become a virtual corpus—as practiced by the Gow and Page edition of the Greek Anthology entitled Hellenistic Epigrams, and continued by their Garland of Philip: here the Greek Anthology is deliberately rearranged by the author, as a series of smaller corpuses. Gow and Page's sizable accompanying commentary also represents the corpuscular force. As a rule, anthologies that grow, like the Greek Anthology, will inevitably become corpuses, because the pressure of selectivity begins to be overwhelmed by the pressure of inclusiveness. The process is obvious in the case of The Norton Anthology of English Literature, with far more being added to each edition than being edited out—because now with each edition the corpus of English literature is successively recognized as being larger than ever.[14]

However, there are some cases that seem to be nearly perfect examples of anthologies, representatives of a canon that seem purely deponent and completely saccharine. Perhaps the purest anthology of them all is William Harmon's Classic Hundred Poems: All-Time Favorites (1998), which Harmon has culled statistically from various editions of The Columbia Granger's Index to Poetry in Anthologies. Here the editor's natural desire, which is to create a corpus, is kept to an absolute minimum, though the statistical method has no way to allow for the fact that editors look above all at what other editors have done in the larger corpus of anthologies, from which they may adopt wholesale, just as the successive editors of the Greek Anthology did (an underestimated, and apparently acceptable, variety of plagiarism). A comparison between the first and second editions of Harmon's Classic Hundred reveals the growing influence of corpus over anthology.[15] The first edition lists the poems in straight order of popularity, thus juxtaposing them in all their variety. The second edition, on the other hand, is arranged chronologically, giving a corpuscular sense of literary history, and moreover contains an additional hundred pages of commentary, notes, glossary, and "further reading," which consists of exactly three scholarly sources for each of the hundred poems (forming the critical corpus).

Another difference between the two editions is that Harmon has used different editions of Granger's Index to derive the hundred greatest hits—and besides reshuffling the order of popularity, all of nineteen poems have been upstaged in the second tally. Here, however, the canon demonstrates its anthological consistency, for the nineteen poems seem to have been very deliberately swapped: for Herbert's "The Collar," in the new hundred there is "The Pulley"; for Wordsworth's "I Wandered Lonely as a Cloud," there is "She Dwelt Among the Untrodden Ways"; for Housman's "Loveliest of Trees, the Cherry Now," there is Burns' "A Red, Red Rose" (debuting this edition all the way up in the nineteenth slot!). A different season, in other words, but all the same things are blooming. The other anthological flowers, the individual authors, are also perennials. Shakespeare, the ultimate canon bellwether in English, has two of his eight poems dropped, but still manages to tie for the "lead" with Donne because Sonnet 129 is now a hit. Other shifts in The Classic Hundred authors are likewise negligible: Keats loses "Ode on a Grecian Urn," but "To Autumn" is now at the very top of the charts. Of course, the analogy that I have been supplying in this paragraph, from the weekly rankings of contemporary music, is faulty: Harmon's list is not based strictly on poems' popularity with fickle readers, as music charts are with listeners (or, strictly, with purchasers and play lists), but on poems' popularity with anthology editors. By comparison, record producers are often far more daring when it comes down to what they are willing to put on an album. Thus it is the creative process of editing an anthology rather than ideas about readers' preferences that should control evaluations of a given anthology's content: how have selections been made, how have they been organized, and how are they contextualized?

In this frame, it must be remembered that an anthology will always begin to become a corpus in the editor's mind. In the preface to the first edition of The Classic Hundred, entitled "The Greatest Hits of Poetry in English," Harmon asks, "Why these? I wish I knew, but I don't" (4). In the preface to the second edition, however, the "greatest hits" trope and its fickleness is gone, for Harmon has become far surer of what he sees:

This is a beautiful description of a corpus, but here Harmon cannot see the anthology for the flowers. This is no corpus, except to one who rereads again and again—what few readers of The Classic Hundred other than the editor are ever likely to do. It must be confessed that only a certain kind of literary piece is favored by an anthology: a shorter, formally controlled poem that imagines loss or devotion, either human or divine. Such are the permillennial epigrams of Meleagar's Garland, and such are the overwhelmingly Renaissance and Romantic lyrics of Harmon's Classic Hundred.

Conclusion: Mixed Canon Pedagogy

Unlike Harmon's Classic Hundred, most canons show more of a balance between the forces of anthology and corpus. Anthological markers include singularity of genre and arrangements that are ranked arbitrarily, as by number or alphabetically. Corpuscular markers include multiplicity of genres and deliberate conglomerations, as by author, theme, or period. (Periods are not anthological; a relisting of a Norton's poems anthologically by their dates would turn literary periods into hash.) As in Meleagar's Garland, most anthologies have some of both, corpuscular groupings within which are jumbled anthological listings. There are also alternations between the forces: the two-volume Nortons are anthologized in shorter one-volume editions; the hefty corpus of the Greek Anthology is winnowed by a series of editors for their contemporary audiences, as Maximus Planudes did, and as almost all of the many English versions of the Greek Anthology have done. On the other hand, no published anthology can fully realize its nominal valence, especially the corpus-sized ones used in literature courses. Any teacher can display these dynamics while still ordering the same "standard anthology" for the lit survey. An ambitious "teaching of the conflict" is one possibility (Graff 1992), but I would also suggest the following two modest alternatives:

1. Teaching the table of contents. Students will naturally tend to approach the table of contents of an anthology from their background as extensive readers: as an index or finding tool for reading assignments. (They will not even bother looking at the contents if a syllabus takes this function and gives them the page numbers.) But an anthology's table of contents is as eloquent as the editorial preface, because to the specialist the table represents not just contents, but a range of deliberate choices—both inclusions and exclusions. Taking a class period to page through and comment on the detailed contents and the other front matter allows students to see how the book they will do most of their reading from has been deliberately shaped, not just gathered (the editor more as floral arranger than flower-picker). The issues to be covered might include: How much space is allotted to each period? Which selections are "new," which "traditional"? Are entire works included (corpus), or just parts of them (anthology)? Are major authors given pride of place with many selections (corpus), and minor authors represented by only one or a few selections (anthology)? Are there groupings based on single genres and themes (anthology), or on whole nations and periods (corpus)? To what extent do introductions and headnotes give interpretations and readings (anthology), or list biographical details and other works by the author (corpus)?

At this point, a general discussion might ensue over what the anthology as a book form might tend to encourage, given its space constraints: More poetry and shorter works? Or, given the need to write time-consuming headnotes, more longer works and more works by the same authors? Braver teachers will then turn to their own syllabi, and evaluate how anthological or how corpuscular their own choices of readings might be. In the context of a seminar or a class research project, an interesting comparison could be made between the contents of the current and earlier editions of the same anthology. By beginning with a reading of the table of contents, the entire course can begin to explore the issue of canonicity for every work: why was this piece selected for this anthology, and also, why are we reading it rather than others? How does this work connect to other works we have read, canonically speaking—how do works measure each other, and what lists or groupings of our own could we make with these works? Here a whole set of "critical thinking" topics emerge for papers and essay exams, with no more than the anthology to go on. Students might even be found (shockingly) going outside of the assigned readings in order to answer the above questions.

2. Adding a corpus to the syllabus. This option provides a dialogic partner for the anthology. While single works by definition will hardly tend to approach the valence of "corpus," many things between two covers do, especially authorial collections of poems or stories. Many such "mini-corpuses" are available in cheap paperbacks, so there is little added cost. There are other simple and inexpensive possibilities as well: having students independently locate in the library and report on, say, the most recent work by an author on the syllabus; adding a novel or a play to one by the same author or one from the same period in the anthology; sampling a collection of an author's nonfiction prose pieces (essays, reviews, criticism, autobiography, letters); reading a "how-to" book or essay by a creative writer or a literary translator; photocopying or showing slides of art or other visual forms that couple with the literature (architecture, emblems, book illustrations, cartoons); playing music or voice recordings that do the same; or showing videos, if this medium is not already being overused. The added corpus, whatever it might be, would function to show how the chosen anthology is not so much defective as deponent, as defined above. The questions that arise from getting a corpus back onto the syllabus may include: Why was this work not included in the anthology? What kind of anthology would it be included in? What genres and themes does the anthology seem to prefer? Do authors act in the same way as anthology editors when making collections of their works? How does the reading of a corpus change the reading of an anthology?

The kind of pedagogy that I am offering here differs in one key way from Gerald Graff's dialogic approach of "teaching the conflicts" (1992), namely that of scale. Graff envisions a dialogism on the level of the entire undergraduate curriculum and thus advocates such practices as team teaching, the clustering of courses, sponsored conferences, and large public forums, all in order to create a dialogue between individual teachers' approaches, which otherwise speak in isolation. Graff suggests that all this does not involve that much more work, since he finds that a dialogue is already latent in the academy, and simply needs to be cross-referenced. But it seems that such ambitions are only possible where programs are willing to put considerable administrative and monetary support behind these impressive dialogic carnivals. My approach here is much more offhand, and may be far more feasible in programs governed by impersonality, a kiss of death for dialogue—especially in larger universities where there may be many graduate or adjunct instructors, and in undergraduate institutions where teaching loads are already heavy.

Instead of asking how one undergraduate course might relate to the other courses that an individual student might possibly take—since even in practice the possibilities are effectively limitless—I am wondering about how a single class period or a single assignment might be tailored to initiate a dialogue within a student. To take the Lewinian force-analysis to its scalar conclusion, there might be a single poem in an anthology that encapsulates the dialogic of the canon, or even a single line or single annotation where all the dynamic factors might be perceptible. If such a dynamic concept can be formed and passed along, the dialogue need no longer be bound by the institutional vagaries of trend-setting programs and the energies of their finite leaders. If a student (speaking now more of a graduate student) possesses a self-consciously dynamic mode of literary concept formation, institutional structures matter little, because the dialogue can now travel anywhere, as the graduate student in particular is bound to do as adjunct or assistant professor. The anthology's dynamic is present in the picking and proffering of a single poetic bloom.

NOTES

[Use BACK button on your browser to return to text.]

[1] A group of case studies of anthologies of English poetry, with comparable conclusions regarding the canonical dynamics of anthologies, appeared after this article was written: see Anne Ferry, Tradition and the Individual Poem: An Inquiry into Anthologies (Stanford: Stanford University Press, 2001).

[2] The "dynamics of canonicity" has also recently been invoked by Robert Alter (2000, 6), whose study of the canon of the Hebrew Bible and its afterlife in modern literature demonstrates that even "a [scriptural] canon can be much more flexible, and less ideologically binding, than prevalent conceptions allow" (5). Alter's emphasis on playful creativity as the dominant force in the history of the canon is meant to soften Harold Bloom's preference (1994) for a canon marked by bouts of Oedipal originality.

[3] For an evaluation of Lewin's importance, see Marrow (1969); Nowak and Vallacher (1998, 12); and especially Ross and Nisbett (1991).

[4] Originally published in the Journal of General Psychology 5 (1931): 141-77, and republished in the 1935 collection A Dynamic Theory of Personality; I refer to the latter. The essay has also been reprinted more recently in Contemporary Psychoanalysis 23 (1987): 517-44, and in Martin Gold, ed., The Complete Social Scientist: A Kurt Lewin Reader (Washington: American Psychological Association, 1999).

[5] On the nature of the two kinds of concepts, Lewin is indebted to Ernst Cassirer's Substanzbegriff und Funktionsbegriff (1910). While Lewin does not, Cassirer does make a historical argument for the succession of these concepts around Renaissance figures such as Galileo (1927).

[6] This strand of thought stems from the "nomothetic" / "idiographic" distinction between the academic disciplines made by Wilhelm Windelband and Heinrich Rickert of the Southwestern or Baden School of Neokantianism.

[7] This solves Newton's second law of motion, acceleration being negligible for glaciers (even galloping ones).

[8] For more on the canon in literary theory, see the special issue of the journal Critical Inquiry, volume 10.1 (1983).

[9] The key classical texts are Pliny, Naturalis Historia 34.55, and Galen, De Placitis Hippocratis et Platonis 5.

[10] See Boardman (1985, 205, pl. 184-85); Pollitt (1990, 75-77); and especially Stewart (1990, 1:160-62, 263-66, 2: pl. 378-82).

[11] This canonicity is emphasized by Galen, De Temperamentis 1.5, Quintilian, Institutia Oratoria 5.12.21, and Cicero, Brutus 296.

[12] Naturalis Historia 34.55. The translation is my own.

[13] The fullest Greek Anthology in English remains the five-volume Loeb edition (Paton 1916-18).

[14] But as the anthologies experience corpuscular weight gain, the anthological force can take counter-measures: some teachers ordering bulky hard-bound collections will have observed their students coming to class with only the pages assigned for the day, anthologies which have been excised in their turn from the textbooks left at home!

[15] The first edition of 1991 was originally entitled The Concise Columbia Book of Poetry, and later reissued as The Classic Hundred: All-Time Favorite Poems. See also Harmon's fuller Top 500 Poems (1992).

WORKS CITED

Abrams, M. H., and Stephen Greenblatt, eds. 2000. The Norton Anthology of English Literature. 7th ed. 6 vols. New York: Norton.

Alter, Robert. 2000. Canon and Creativity: Modern Writing and the Authority of Scripture. New Haven: Yale University Press.

Barr, James. 1983. Holy Scripture: Canon, Authority, Criticism. Philadelphia: Westminster Press.

Bloom, Harold. 1994. The Western Canon: The Books and School of the Ages. New York: Harcourt Brace.

Bloom, Lynn Z. 1999. "The Essay Canon." College English 61: 401-30.

Boardman, John. 1985. Greek Sculpture: The Classical Period. London: Thames and Hudson.

Cameron, Alan. 1993. The Greek Anthology: From Meleager to Planudes. Oxford: Clarendon Press.

—. 1996. "Anthology." Pp. 101-02 in The Oxford Classical Dictionary, ed. Simon Hornblower and Antony Spawforth. Oxford: Oxford University Press.

Cassirer, Ernst. 1910. Substanzbegriff und Funktionsbegriff: Untersuchungen Ÿber die Grundfagen der Erkenntnistkritik. Berlin: B. Cassirer.

—. 1927. Individuum und Kosmos in der Philosophie der Renaissance. Leipzig: B. G. Teubner.

Damrosch, David, ed. 1999. The Longman Anthology of British Literature. 6 vols. New York: Longman.

Fowler, Alistair. 1979. "Genre and the Literary Canon." New Literary History 11: 97-119.

Gow, A. S. F., and D. L. Page, eds. 1965. The Greek Anthology: Hellenistic Epigrams. 2 vols. Cambridge: Cambridge University Press.

—. 1968. The Greek Anthology: The Garland of Philip, and Some Contemporary Epigrams. 2 vols. London: Cambridge University Press.

Graff, Gerald. 1992. Beyond the Culture Wars: How Teaching the Conflicts Can Revitalize American Education. New York: W. W. Norton.

Gutzwiller, Kathryn J. 1998. Poetic Garlands: Hellenistic Epigrams in Context. Hellenistic Culture and Society, vol. 28. Berkeley: University of California Press.

Harmon, William, ed. 1992. The Top 500 Poems. New York: Columbia University Press.

—. 1998. The Classic Hundred Poems: All-Time Favorites. 2nd ed. New York: Columbia University Press.

Harris, Wendell V. 1991. "Canonicity." PMLA 106: 110-21.

Hayles, N. Katherine. 1984. The Cosmic Web: Scientific Field Models and Literary Strategies in the Twentieth Century. Ithaca: Cornell University Press.

Leeper, Robert. 1943. Lewin's Topological and Vector Psychology: A Digest and a Critique. University of Oregon Monographs: Studies in Psychology. Eugene: University of Oregon.

Levin, Harry. 1981. "Core, Canon, Curriculum." College English 43: 352-62.

Lewin, Kurt. 1935. "The Conflict between the Aristotelian and Galileian Modes of Thought in Contemporary Psychology." Pp. 1-42 in A Dynamic Theory of Personality: Selected Papers, trans. Donald K. Adams and Karl E. Zener. New York: McGraw-Hill.

—. 1951. Field Theory in Social Science: Selected Theoretical Papers. Ed. Dorwin Cartwright. New York: Harper.

—. 1936. Principles of Topological Psychology. Trans. Fritz Heider and Grace M. Heider. McGraw-Hill Publications in Psychology. New York: McGraw-Hill.

Marrow, Alfred J. 1969. The Practical Theorist: The Life and Work of Kurt Lewin. New York: Basic Books.

Mey, Harald. 1972. Field-Theory: A Study of Its Application in the Social Sciences. Trans. Douglas Scott. London: Routledge; New York: St. Martin's Press.

Minnis, A. J. 1988. Medieval Theory of Authorship: Scholastic Literary Attitudes in the Later Middle Ages. 2nd ed. Middle Ages Series. Philadelphia: University of Pennsylvania Press.

Natoli, Joseph P., ed. 1987. Tracing Literary Theory. Urbana: University of Illinois Press.

Nowak, Andrzej, and Robin R. Vallacher. 1998. Dynamical Social Psychology. New York: Guilford Press.

Paton, W. R., trans. 1916-18. The Greek Anthology. 5 vols. Loeb Classical Library. Cambridge: Harvard University Press; London: W. Heinemann.

Pfeiffer, Rudolf. 1968. History of Classical Scholarship from the Beginnings to the End of the Hellenistic Age. Oxford: Clarendon Press.

Pollitt, J. J. 1990. The Art of Ancient Greece: Sources and Documents. Rev. ed. Cambridge: Cambridge University Press.

Ridgway, Brunilde Sismondo. 1981. Fifth-Century Styles in Greek Sculpture. Princeton: Princeton University Press.

Ross, Lee, and Richard E. Nisbett. 1991. The Person and the Situation: Perspectives of Social Psychology. Philadelphia: Temple University Press.

Stewart, Andrew F. 1990. Greek Sculpture: An Exploration. 2 vols. New Haven: Yale University Press.

Todorov, Tzvetan. 1984 . "A Dialogic Criticism?" Raritan 4.1: 64-76.

van der Veen, C. J. 1999. Fundamentals of Glacier Dynamics. Rotterdam: Balkema.

**