CHIN 403                                                                    Essay 1                                                                25 May, 2005

Ray Brownrigg


The Origins of Xiqu, or Chinese Opera


Abstract: This essay investigates the origins of xiqu, or Chinese opera, and proposes that nanxi, or southern theatre, was the earliest form of xiqu.  This proposal is analysed from two perspectives.  Firstly the claim that nanxi was indeed an early form of xiqu is investigated, and then the claim that nanxi was the earliest of the possible candidates is discussed.  Finally, some discussion is presented of the political and cultural environment which may have contributed to the development of true theatre at this time and place.



One difficulty that arises when discussing traditional Chinese opera using the English language is the difficulty in choosing an adequate translation for the Chinese word xiqu (戏曲).  Various translations include “theatre of song”, “music drama” and “opera”, but none of these really conveys the correct meaning, and most attempts will confuse an English-speaker who is already familiar with Western meanings associated with the words theatre, music, opera, drama and play.  Most Westerners, though, have some concept of “Chinese opera” (most commonly defined historically by “Peking Opera”) as being quite different from any of these.  For this reason, the term “Chinese opera” is a good start for a translation of xiqu.


Some of the components of Chinese opera had existed for many centuries, or even several millennia, before they all coalesced into what is considered to be true xiqu.  There exists evidence of various forms of singing, dancing, acrobatics, story-telling, humour, acting, makeup and costumery in early written records, in some cases as early as the second millennium BC.  But just when these components did all coalesce into a genuine form of xiqu is the subject of much discussion and continuing research, and is by no means a universally accepted thesis.  Not only is there a difficulty in identifying which of the contenders did satisfy all the criteria of true xiqu, there is also some discussion and room for further research in identifying, of those contenders that are actually considered to have been true xiqu, just when each first arose in its true xiqu form.  Part of the problem here is that for most, if not all, contenders, the nature of the theatre form was continuously evolving.  Thus while it may well be the situation that a particular performance or a particular script could mark the transition point to “true xiqu”, identifying the exact timing of such a performance or script is somewhat futile from such a distance in time.


It has already been mentioned that many components of sight, sound, action and emotion can be manifest in a xiqu performance, but while every xiqu performance will contain some or all of these components, the definition of true xiqu goes beyond the performance itself.  Mackerras (1990, p8) defines drama in its widest sense as “a branch of the performing arts in which each item has a story and, in formal performances, one or more performers who impersonate and dress themselves as characters in such a way as to portray relatively complex interrelationships between or among these characters.”  Thus Mackerras considered the existence of a plot, characterisation and the use of costumes to be distinguishing components of drama.  Xiqu nowadays seems to be defined more by example (such as “Peking Opera”) than by formal definition.  When a definition is offered, it tends to be even more vague, such as “traditional Chinese opera” or “indigenous Chinese opera”.  Unfortunately, these do not give us enough detail for us to be able to identify what in the past was or was not true xiqu.  Now since a prerequisite to being able to determine the earliest xiqu form is to have available a definition of true xiqu, we must formulate such a definition in such a way as to be able to measure early theatrical forms against this definition.


Finally, to conclude this introduction, we must consider which are the possible theatrical forms that would be in contention for consideration as the earliest xiqu form.  The precursors of traditional Chinese theatre go back several millennia, but it is not until towards the end of the first millennium AD, namely from the late Tang dynasty (618-907) through the Five Dynasties period (907-960) to the early Song dynasty (960-1278) that serious contenders started to flourish. These were the canjun xi ( 参军戏), or adjutant play, from the mid-Tang to early Song, the Song zaju (宋杂居) and Jin yuanben (金原本) plays from the Northern Song and Jin dynasties respectively, nanxi (南戏) or southern drama from the late Song dynasty onwards, and Yuan zaju (元杂剧) from the early Yuan onwards.  Since we are proposing that nanxi was in fact the earliest xiqu form, we must look at each of the above contenders and measure them against our definition.


How do we define true xiqu?

From the current research literature, some of the considerations of what constitutes true xiqu seem to be:

  • Was there a story with a plot (rather than just a description of what happened)?

This is mentioned above as one of Mackerras’ definitions of the requirement of drama in general and thus must also be considered to be required for xiqu.

  • Was there a cast of characters (rather than one or more narrators)?

This is how we can ensure that the performers “impersonate characters”, being also one of Mackerras’ requirements.

  • Was there make-up (or masks) and costumery?

This is the third of Mackerras’ requirements, namely that the performers “dress up”.

  • Was there a script (perhaps also with stage directions)?

The existence of a script has long been considered to be a prerequisite for a xiqu performance and seems quite a reasonable requirement, since the lack of a script raises questions about the existence of a fixed plot.

  • Was there a mixture of dialogue, action and song?

The amalgamation of multiple talents is one thing that distinguishes xiqu from Western performing arts (where opera usually has no action, and drama usually has no singing).

  • Was there music (not only just to accompany the singing)?

Again, musical accompaniment to all parts of a performance is what distinguishes xiqu from Western performing arts.

  • Was there a well-defined role-category for each character portrayed?

The use of role categories was well-entrenched in Chinese performing arts before the rise of xiqu, and has persisted as a crucial component of xiqu.

  • Was the performance the principal (or only) feature of the production in which it appeared?

This is perhaps the most difficult requirement to test, but is nevertheless important, since it identifies such a performance as the primary purpose for the audience to gather, and so indicates that the performance is important in its own right, rather than as a side-show or prelude or entr’acte for some other more important performance.


It seems reasonable to accept that if all the above requirements were satisfied, there would be no question that the theatre form being considered was indeed true xiqu.  Where difficulties arise is if most of the requirements are satisfied, but some are missing.  If the lack was occasional, then this would not necessarily disqualify the contender, but a consistent lack of any one of the above would result in something that could not be considered to be true xiqu.


Was nanxi true xiqu?

Now we must measure nanxi up against the above stated criteria, to determine that it is indeed a genuine xiqu.

  • Was there a story with a plot?  The original scripts of several nanxi plays, discovered in 1920 (Sun 1996, 142), one of which is considered to be written during the Southern Song dynasty, serve to provide evidence that nanxi plays did indeed develop dramatic plots.  Many of the nanxi plays had romance as their theme (Mackerras 1990, 28; Dolby 1983, 33), which would need a plot to become successful.
  • Was there a cast of characters?  The use of role categories to identify characters within nanxi scripts (Sun 1996, 150) certainly provides evidence that there was a cast of characters, but the presence in the scripts of stage directions involving multiple characters (Sun 1996, 154) reinforces the notion that there were multiple performers.
  • Was there make-up (or masks) and costumery?  The nanxi play “First Place Scholar Zhang Xie”, perhaps the earliest known nanxi script, contains several references to the makeup worn by characters within the dialogue.  While this evidence is scarce indeed, there is certainly no evidence to the contrary.
  • Was there a script?  The discovery of several nanxi play scripts in 1920 and the subsequent research which has gathered together many fragments of nanxi scripts, provides direct evidence that in fact nanxi plays were performed from scripts.
  • Was there a mixture of dialogue, action and song?  The scripts and fragments of scripts provide significant evidence of a mixture of action, dialogue and singing within nanxi plays.
  • Was there music?  There is little information available about the use of musical instruments during nanxi plays, but what information there is, from the introductions provided in scripts and, more rarely, a musical score (Sun 1996, 149), indicates that performances were indeed accompanied by at least percussion instruments.
  • Was there a well-defined role-category for each character portrayed?  Role categories are well known within nanxi plays, the strongest evidence being from the use of role category names rather than character names to refer to the different characters in nanxi scripts.
  • Was the performance the principal (or only) feature of the production in which it appeared?  While the actual length of the small number of nanxi scripts indicates that these are indeed full-length productions, other evidence is indirect in the sense that nanxi was considered to be a competitor to Yuan zaju, and a forerunner of the later chuanqi which arose during the late Yuan and early Ming dynasties.


It is difficult not to notice that much of the evidence cited immediately above comes from the content of known nanxi scripts.  Thus it is easy to understand why early scholars did not consider nanxi to be a true form of xiqu until after the rediscovery in 1920 of original play scripts.


How do the other contenders measure up?

The earliest contender for the title of true xiqu is canjun xi, also known as adjutant plays, which developed at least from the 8th century, during the late Tang dynasty, and flourished into the early, or perhaps even mid-, Song dynasty (Mackerras 1990, 23; Dolby 1983, 13).   This theatre form is well known for its introduction of role categories into traditional Chinese theatre and, while it satisfies many of the requirements for being true xiqu, there are a few important requirements which are missing.  Canjun xi can best be described as short skits, and as such do not satisfy the requirement that a xiqu performance constitute a full-length production, or be based on a complex plot.  Further because of the absence of, perhaps just still undiscovered, scripts, this theatre form cannot be considered to be true xiqu.


During the Northern Song dynasty, canjun xi tended to be replaced by what is now known as Song zaju which was similar, but was distinguishable by the appearance of more role categories.  One improvement towards genuine xiqu status that Song zaju achieved over canjun xi was the appearance of performances that were of significant duration (Mackerras 1990, 25).   While this was a step in the right direction, it remained the case that such extended performances were merely collections of largely unrelated small items, much more like a variety show than an extended dramatic performance.  Thus none of the missing requirements was added, and so Song zaju remains like canjun xi as a forerunner to, but not a true, xiqu.  Similarly Jin yuanben, which flourished during the Jin dynasty in the North, evolved from canjun xi, and perhaps replaced Song zaju, but added no features which would allow one to consider its status as being any closer to true xiqu.


The only other possible theatre form which might be considered to be the earliest xiqu form is Yuan zaju, which arose with the Yuan dynasty, immediately following the Northern Song.  Like nanxi, the characteristics of Yuan zaju are preserved in the presence of original scripts.  In addition, Yuan zaju enjoys the existence of much more commentary to reinforce what information the scripts provide, as well as a surviving wall painting which provides useful information about an actual performance.  Now while there were distinctive differences between the nanxi and Yuan zaju forms, the existence of the play scripts and the other written material provide sufficient evidence for Yuan zaju to be considered a true xiqu form.  Thus it now remains to determine which of nanxi and Yuan zaju arose the earlier.


Was nanxi earlier than Yuan zaju?

Although in general the performing arts in China during the early part of the second millennium were somewhat independent of changes of government, the rise of Yuan zaju seems to be an exception.  The Mongol conquest of the Jin dynasty in the North in 1234 and the subsequent suppression of scholars, by way of the abolition of the Civil Service Examinations from 1237, provided the catalyst for the birth of the xiqu form known as Yuan zaju.  This then identifies reasonably precisely the timing of the rise of Yuan zaju as being soon after the demise of the Jin dynasty, during the Mongol rule period preceding the formation of the Yuan dynasty.  The rise of nanxi, on the other hand, is a much more difficult target to identify, mainly through the lack of recorded information.  However it seems clear that nanxi predates Yuan zaju by at least several decades, and probably by more than 100 years (Sun 2002).


Some political, economic and cultural influences

Apart from the influence of the rise to power of the Mongol Yuan dynasty mentioned above, political influences on traditional Chinese opera have tended to be somewhat indirect.  For example the overthrows by both the Jin and Yuan dynasties pushed refugees south to the more stable Southern Song dynasty and this would have resulted in much greater concentrations in population in the areas around Hangzhou, where it seems that nanxi arose.  Economically, South China was thriving anyway, and the influx of a greater population would have produced far greater potential audiences which in turn would have made the existence of troupes of performers much more economically viable.  Further, because of the prosperous economy, there was progressive urbanisation and a general population mobility, both of these in addition to the normal movements of traders, all of which could lead to the spread of theatre forms.    Finally, this was during a period of highly developed culture, which may well have engendered a general movement towards the performing arts as a means of entertainment.



Evidence points to nanxi, which arose around the time of the move to the South of the Song dynasty, as being the earliest appearing full xiqu form.  While information about the timing of the rise of nanxi is scarce, there is sufficient evidence to place it significantly before its nearest rival, Yuan zaju.  We have discussed how nanxi does satisfy the conditions to be a full form of xiqu, and how it is the earliest appearing such.  While there is little written evidence apart from the nanxi scripts themselves, those scripts provide enough evidence to make this claim. 



Chen, Jack, The Chinese Theatre, (London: Dennis Dobson Ltd, 1949)

Dolby, William, “Early Chinese Plays and Theatre,” in Chinese Theatre from its Origins to the Present Day, ed. Colin Mackerras, (Honolulu: U Hawaii P, 1983), pp 7-31

Dolby, William, “Yuan Drama,” in Chinese Theatre from its Origins to the Present Day, ed. Colin Mackerras, (Honolulu: U Hawaii P, 1983), pp 32-59

Mackerras, Colin, Chinese Drama, A Historical Survey, (Beijing: New World P, 1990)

Sun Mei, “Performances of Nanxi”, Asian Theatre Journal, 13:2 (Fall 1996), pp 141-166

Sun Mei, “Exploring the Historical Development of Nanxi, Southern Theatre”, CHINOPERL Papers, 24 (2002), pp 35-65.