Review by
Edward Slingerland
Philosophy East and West, v50 #1 (January 2000) 137-141

The text of this review article is taken from Thomas Carlson's web site, where it was originally included with the permission of the author.

Why Philosophy is Not "Extra" in Studying the Analects

The Original Analects by E. Bruce Brooks and A. Taeko Brooks is an extraordinary book in many ways, and is clearly required for anyone concerned with early Confucian thought. While it fails to substantiate most of its more radical claims, it performs an invaluable service by forcing scholars of the text to take seriously the fact that the Analects is a heterogeneous collection of writings representing different time periods and a variety of concerns - a fact that has often merely been given lip service by students of Confucianism, this reviewer included.

Unfortunately, few readers will find it a readily accessible book. The style is reminiscent of the late Peter Boodberg: the translation is awkward yet etymologically precise, careful attention is paid to sinological detail, and there is marked fondness for neologisms and idiosyncratic scholarly conventions. Some of these conventions (the method of citation, for instance) represent a slight improvement over standard practice, but many (such as the romanization and dating) are off-putting and unhelpful. It is not at all clear, for example, why it is less "culturally parochial" (p.2) to indicate a date's relationship to the birth of Christ by preceding it with the number "0" rather than following it with the letters "B.C.E."; this arbitrary decision simply creates another stylistic barrier for the reader to overcome.

The Brookses argue for a radical reorganization of the chapter order in the received text, and identify large numbers of interpolations within each individual chapter. They also claim that the text was composed over a much longer period of time than has generally been accepted - the later strata, in their account, being put together as late as the third century B.C.E. If correct, this would have significant philosophical implications. This is illustrated most dramatically in Appendix 5, where they demonstrate that placing chapters 1-3 before chapter 4 (identified by them as the earliest chapter) significantly alters the manner in which these chapters are interpreted. For instance, it is proposed that an original emphasis on public, political virtue can be seen as shifting to an emphasis on the personal and the domestic (p. 312), and it is argued that an original focus on public effectiveness was warped by a later obsession with scholastic "learning" or xue - the term xue itself shifting in meaning from "imitation/emulation" (xiao) of politically active models to mere book learning (p.27). The Brookses similarly identify what they claim to be a shift in the meaning of ren from a martial virtue to a ritualistic one (p.254), resulting from an increased emphasis on ritual practice (li) by the later members of the Lu Confucian school. "The debate on whether ren or li is central to Confucianism is thus solved," they proclaim. "Ren is central to Confucius, whereas li is central to Confucianism" (p.16).

The power of this argument is somewhat weakened by the fact that it is largely circular. For example, an exclusive focus on public virtue in chapter 4 is achieved only by eliminating passages 18-25 of the received text, and the main rationale for eliminating these passages is in turn the fact that they "emphasize domestic and personal virtues, in contrast with the official focus of 4:1-17, and were presumably added to legitimize a later doctrinal shift in that direction" (p.208). Indeed, much of the Brookses' micro-periodization of the text (the dating of individual chapters) is similarly speculative and circular in nature, while their macro-periodization (their overall chronological framework) also rests upon quite shaky ground. With the exception of their treatment of chapter 4, they very seldom cite hard internal linguistic or historical evidence to support their dating; the only firm dates that can be gleaned from the entire argument are the use of a posthumous epithet for Ai-Gung in 6.3 and the reference to the death of Zengzi in 8.3, which indicates that these passages must have been written after 469 B.C.E. and 436 B.C.E., respectively (p.205). More usually, an actual historical reference from the text is taken to be a coded reference to a much later event, and this later event is sometimes turned around and used as a "proof" for another date further on in the book. While such metaphorical uses of historical accounts are, of course, quite common in early China, it would seem inadvisable to rely upon such a possibility as "proof" for one's dating scheme.

The Brookses' metaphorical interpretations of historical references are really only convincing if one has already accepted their macro-periodization scheme. This scheme - the bedrock assumption upon which the fragile edifice of further assumption is built - is the so-called "accretion theory": the theory that the text of the Analects was composed at a constant rate, in distinct bands each one chapter thick. Added to this assumption is the claim that the fragmentary nature of chapter 20 of the received Lu version of the text can best be explained by assuming that its accretion was cut off by the Chu conquest and the absorption of Lu in 249 B.C.E. Arguing that the text did not begin to be composed until after Confucius' death in 479 B.C.E. (based on the fact that his sayings are always introduced by a third-person formula), this gives a 230-year period over which the text was "accreted". By further comparing the Lu version of the Analects with the Qi version - which was reportedly two chapters longer than the Lu version and whose compilation was similarly interrupted by the Chu conquest of Qi in 221 B.C.E. - the Brookses arrive at a "rate of accretion" of 12.7 years per chapter, or two chapters per generation. Beginning with chapter 4 (which they identify as the earliest stratum of the text, based on some fairly convincing internal linguistic evidence), every pair of two chapters is then identified with a specific successor to the headship of the Lu branch of the school of Confucius.

This accretion theory is an excellent example of a metaphor from the natural sciences being inappropriately applied to the humanities. Texts are not like redwood trees or coral reefs, and there is no reason for us to think that the Analects was not composed in a fairly haphazard manner. It may be the case that chapters or parts of chapters were added in groups to a multi-chapter original core, or that the text simply represents an anthology of early Confucian writings from different time periods that was cobbled together by a single author at a very specific date. Similarly, the fragmentary and extremely heterogeneous nature of chapter 20 of the received text suggests to me that it represents a dustbin for material that could not be worked into other chapters, rather than the "stump" of a chapter whose growth had been violently interrupted. Furthermore, although the third-person nature of the Master's sayings suggests that he himself did not compose the text, there is no reason at all to see this as "proof" that the Master was no longer among the living when the text was composed. In fact, we would be rather surprised to find such an early text being composed in the third person.

Additionally, the Brookses often find evidence for their dating scheme in passages that are supposedly responses to the Mohists, Legalists, or Daoists, or to the Mencian and Xunzian branches of Confucianism. One is then led to wonder why these rival thinkers are never mentioned by name, as was the standard practice by the time of the Mencius and the Xunzi. Going beyond such obvious objections, one could further argue that the Brookses' narrowly sinological focus blinds them to some of the deeper philosophical problems with their account. That is, their investment in their elaborate Just-So chronology, their fascination with sinological arcana, and their creativity in identifying "coded" references to events and personages causes them to overlook a set of larger philosophical issues that contradict their dating scheme. For instance, although it is claimed that chapters 12 and 13 were written either by Mencius himself or under his direct influence, we in fact do not find anywhere in the Analects even a dim awareness of the highly developed conceptions of human nature that formed the basis of the debate between Mencius and Xunzi. We similarly look in vain for evidence that any of the authors of the Analects were familiar with the elaborate psycho-physiological techniques of qi cultivation advocated in the "Inward Training" and "Techniques of the Mind" chapters of the Guanzi - techniques that apparently exerted such influence on subsequent thought that the authors of the Mencius, Zhuangzi, and Xunzi had no choice but to appropriate them. We do, indeed, begin to see hints of a larger concern with qi and human nature in (respectively) Analects 16.7 and 17.2, but in comparison with the Guanzi or Mencius such late passages are philosophically quite primitive.

There is also nowhere in the Analects a trace of the more specific conception of the heart/mind (xin) that is assumed in late Warring States texts - the picture of the heart/mind as the locus of discrimination (bian), reasoning, and language use - and the authors of the Analects were apparently also unaware of the later Mohist theories of language and logic that no late Warring States author could afford to ignore. The internal philosophical and terminological evidence would thus suggest that no portion of the Analects was written after the early fourth century B.C.E. We could thus take the text as a whole as representing the state of the early Confucian school as it existed prior to the philosophical explosion of the "Hundred Schools" period and the Mencian innovations demanded by this increasingly sophisticated intellectual milieu, and this in turn allows us to hold a view of the text and its place in pre-Qin Chinese thought not terribly different from that dictated by the traditional chronology.

The Brookses' lack of philosophical orientation reveals itself as well in their resolutely political rendering of the text. Practicing a rather extreme hermeneutics of suspicion, they systematically discount the possibility that philosophical developments of early themes by later compilers might actually be genuine attempts to elucidate the Master's teachings and make the Analects relevant to a new age. Rather, doctrinal innovations are generally seen in terms of political stratagems designed to enhance the prestige of one line of disciples over another. A similar politicization is revealed in much of the commentary to the text. For instance, the comment on Analects 1.3 (p.146) assumes that the dislike for "glibness" expressed in this passage is merely a peevish slander aimed at those more successful than the Lu Confucians in gaining an official hearing; the possibility that the author of this passage has a philosophical reason for being suspicious of facile talkers is never explored.

On the other hand, the Brookses are on the whole to be commended for the commentary that accompanies their translation, for it represents a gold mine of sinological detail and often provides some quite original insights into the text. For instance, the description of Japanese gagaku in the commentary of 3.23 (p.85) makes this passage come alive to the reader, while at the same time rendering its message considerably more intelligible than any other interpretation I have read. Indeed, the issue of its periodization scheme aside, The Original Analects represents an invaluable sourcebook for students of the text. It allows them entrée into a rarified scholarly atmosphere that cannot but enrich and deepen even the specialist's understanding of the text, and this is perhaps its greatest virtue. Whatever one may in the end think of the validity of the arguments made by the Brookses in support of their accretion theory, one cannot help but be impressed by the immense erudition and linguistic acumen expended in their effort.

To summarize, then, if we eliminate the more speculative and unsubstantiated elements of the Brookses' argument, it would seem that we have not advanced much beyond the periodization scheme proposed by Cui Shu (1740-1816): chapters 1-10 represent an early core to which 11-15 and (last of all) 16-20 were added. [Note: Some scholars, such as Arthur Waley, have suggested that chapters 3-9 represent the original core, arguing that chapters 1-2 lack the thematic unity of 3-9 and that chapter 10 is "irrelevant." I would perhaps grant the first point, but agree with Ito Jinsai (1627-1705), whose views are reported by the Brookses on p. 201, that chapter 10 - a portrait of Confucius, the perfected ritual master, in action - serves a crucial role as a thematic "cap" for the chapters that precede it.] The Brookses are no doubt correct in identifying later interpolations within the earliest stratum, and the archaic linguistic features they identify in chapter 4 may indeed indicate that chapters 1-10 do not represent a strictly chronological progression. In the final analysis, however, we might have to admit that a definitive micro-periodization of every chapter of the Analects is something that may forever remain beyond our ability to realize. The power of the Brookses' effort derives, then, not so much from their specific conclusions, but rather from their success in shaking to the foundations our traditional conception of the Analects and forcing us to consider seriously the philosophical difficulties in approaching such a heterogeneous and stratified text.

Edward Slingerland
University of Southern California

[For our response, click HERE]

 Back to Responses Index Page

20 March 2006 / Contact The Project / Exit to Publications Page