ONE    New Criteria for “Good” Medical Formulas

CLAIMING EFFECTS WITHOUT WITNESSING

Throughout the history of imperial China, claiming that recorded remedies had the desired effects had been a significant strategy for building the credibility of a medical treatise. In ancient China, when medical treatises were often attributed to legendary and divine figures, such claims were made by associating the transmission of remedies with those prominent figures. Starting in the second century, medical authorship shifted away from this kind of attribution. Medical writers began to identify their names as authors of treatises and sometimes even composed prefaces to express their intentions.¹ Claiming effects for remedies collected in a treatise thenceforth became a way to generate and bolster specific authors’ reputations. The primary method of making such a claim shifted from creating a connection with a legendary figure to including the character yan in a formula’s record. Yan, depending on the context, could mean “verify” as a verb, “effective” or “efficacious” as an adjective, or “effect” or “verification” as a noun. It would appear in the names of formulas, such as “a formula for treating foot weakness with magnificent effects,” or at the end of a formula record, in phrases such as “having effects,” “frequently being effective,” “using it [i.e., the given formula] and achieving remarkable effects,” “having good effects,” and so forth.

The inclusion of yan in a medical formula seems to have been less a formulaic expression than an empirical proof.² The majority of medical authors between the second and ninth centuries, however, did not specify who witnessed the effects. Only a handful described witnesses who had attested to the effects of some of the formulas. These references included their names and personal backgrounds, as well as the dates when the formulas were applied.³ According to modern scholarship, the first formulary that incorporated authors’ medical cases in China is Essential Formulas Worth a Thousand in Gold, for Emergency Preparedness (Beiji qianjin yaofang, hereafter Essential Formulas), which was completed by the eminent physician Sun Simiao (ca. 581–682), sometime before 659. Sun included twenty-five of his medical cases, involving a variety of remedies, in Essential Formulas. Using these to support the efficacy of the formulas he recorded, Sun exhibited an innovative empirical strategy that was rooted firmly in personal experience.⁴

Other examples of formulas incorporating witnesses’ information appear in Mr. Cui’s Collections of Essential Formulas (Cuishi zuanyaogfang), a treatise presumably compiled by a high-ranking civil official named Cui Zhiti (?–681).⁵ Mr. Cui, for example, recorded how Cui designed a drug remedy to treat diarrhea and cured a dozen people.⁶ It also detailed an experience in which a monk assisted with the delivery of an infant, a record that was transcribed from another earlier but now lost treatise, The Regulating-Qi Formulary (Tiaoqi fang). The treatise is attributed to a renowned monk, Tanluan, who was active during the Northern Wei dynasty (386–534). In the record, Tanluan used the first-person pronoun to document vividly how he helped Mr. Yang’s daughter-in-law deliver a child without incident.⁷

Granted, no reader would doubt that some portion of the effects claimed in formularies for certain medicines derived from the authors’ own witnessing or hands-on experience. Authors before the ninth century generally did not, however, identify individuals who testified to the effects of a formula. Sun Simiao’s Essential Formulas is an example of this practice. Of some 4,000 remedies collected in this formulary, Sun only incorporated twenty-five medical cases. That ratio (0.6%) is surprisingly low, especially when we consider that Sun, as an active and eminent physician of his time, presumably attended a considerable number of healing events. The twenty-five cases must constitute only a small fraction of such events in which he was involved. Sun did not explain the rationale behind his selection of these few cases. The low ratio and the absence of an explanation together imply that in Sun’s view, the documentation of those who witnessed the effects of a formula was not instrumental to building a remedy for its reliability.⁸

The scarcity of surviving formularies that were completed before the Song dynasty might be the chief cause of the paucity of records that describe individuals’ witnessing the effects of certain formulas. Except for Sun Simiao’s two, we have only fragments of the contents of the majority of pre-Song formularies, which are preserved in two large-scale encyclopedic formularies. One is The Formulary of Secrets and Essentials from the Imperial Library (Waitai miyao fang, hereafter The Imperial Library Formulary), which Wang Tao (702–772) compiled and presented to the Tang court in 752.⁹ Another is The Formulary at the Heart of Medicine (Ishimpō), which the court physician, Tamba Yasuyori (912–995), compiled between 982 and 984 in Japan and presented to the court in 986.¹⁰ The fragments of information left by those incomplete pre-Song formularies prevents scholars from properly analyzing the portions that explain medical formulas for which there were witnesses who testified to their effects.

In addition to the names of medical formulas, the character yan (verify, being effective/efficacious, effect or verification) also appears in formulary titles, where it generally immediately precedes another character, fang (methods, formulas, or formularies), as in “collections of effective formulas.” Literary scholars and historians have long analyzed book titles when exploring paratexts in European books. Literary scholarship observes that they serve two descriptive functions. One is “thematic,” whereby a title signifies the contents, themes, and topics of a given book. Another is “rhematic,” indicating its narrative form or genre.¹¹ Recent scholarship has applied this observation in examining titles of medical publications in imperial China, showing that when the word fang appeared in formulary titles, it was used either thematically or rhematically. When it was used thematically, the word should be translated as “formulas” (designating the main content of a book); when it was used rhematically, it should be translated as “formularies” (designating a textual genre that contains formulas and other remedies).¹²

Formulary titles containing the phrase “effective formulas” (yanfang) began proliferating in the Era of Division (220–589) and continued to be popular during the Song dynasty. The use of the phrase “effective formulas” in book titles denotes the thematic, not rhematic, qualities of a formulary, because the phrase implies that the remedies collected in a medical treatise were tested and proven effective. Given the fragmentary nature of extant pre-Song formularies, however, it is difficult to examine the degree to which their authors narrated scenarios in which allegedly “effective formulas” achieved their therapeutic objectives. A formulary that gives us a sense of such fragments is The Formulary that Treats Cold Damage Disorders and Is Bodily Verified (Liao shanghan shen yan fang). The title suggests that this formulary consisted mainly of formulas whose effects had been attested to, based on the author’s personal experience with them. Unfortunately, the treatise is now entirely lost and known only by its title, as listed in the imperial bibliography of Sui Dynasty History (Sui shu, compiled by Tang court scholars from 621 to 636).¹³

It is not until the ninth century that we find a formulary that highlighted witnesses to the effects of a formula. The text is titled Passing on Trustworthy Formulas (Chuanxin fang, 818) and was written by a civil official, Liu Yuxi (772–842). In his preface to this formulary, Liu indicated that he had verified the effects of all the formulas he collected. He recalled that, in 805, when he was exiled to Lian Prefecture (in Guangdong), the governor of Jianghua County (in Hunan) and his friend, Xue Jinghui, gave him A Collection of Effective Formulas from Past to Present (Gujin jiyanfang), which Xue had compiled, and hoped Liu would “complement what this formulary lacks.” In response, Liu thereby selected approximately fifty formulas from his box that he had tested. Liu claimed that “all of them have grounds and this is why I used Passing on Trustworthy as the title.”¹⁴ Although Liu did not specify whether the character zi (grounds) referred to the textual origins of those formulas or to sources of information about their effects, the narrative purpose of his compilation favors the latter interpretation. Along this line, the word xin (trustworthy) in the formulary title refers to the quality of the sources of information about the formulas’ effects.

Moving to the Song dynasty, we find a similar emphasis on the reliability of information sources regarding formulas in Mr. Wang’s Formulary for Broad Relief (Wangshi boji fang). Its author, Wang Gun, a low-ranking official who was responsible for selling alcohol in Qiantang (in Zhejiang), finished this treatise in 1047. Wang said that he had devoted himself to collecting formulas for two decades, in the process acquiring seven thousand of them. “Those formulas,” he declared, “are all transmitted from family documents and obtained from relatives and old friends. They are not received through eavesdropping or oral transmission.”¹⁵ Of the seven thousand formulas, Wang selected those that were particularly “essential” (jingyao) and known to be effective, compiling them into Broad Relief.

The empirical strategy emerged in the seventh century, marking a new persuasion strategy in the formulary genre that foregrounded knowledge rooted in authors’ personal experience. With few exceptions, however, the authors of formularies that were completed between the seventh and ninth centuries apparently applied the empirical strategy only rarely. They did not regard the inclusion of information identifying who verified a given formula’s effects as central to legitimating the reliability of their work. Most of these authors believed that simply declaring that their medical formulas had been used and had proven effective sufficed to establish their reliability. There was no need to document their efficacy by referring to actual cases or patients. I use the term “effective-formula strategy” here in reference to this single means of reliance on formulas known to be effective, thereby distinguishing this strategy from the empirical strategy that Shen Kuo followed in his Good Formulas.

NEW CRITERIA FOR THE GOODNESS OF FORMULAS

Questioning the evidential value of the effective-formula strategy, Shen Kuo in his preface to Good Formulas proposed various elements of a qualifying and trustworthy formula. One such element is documenting witnesses: “What I call ‘good formulas’ have to be those of which I have witnessed their effects and then written them down. Hearsay has no role in my identification.”¹⁶

Scholars have regarded this sentence as a path-breaking event in formulary writing and often cite it when examining Shen’s empiricism and, by extension, exploring that of Song scholar-officials more generally.¹⁷ But what did Shen mean by the seemingly straightforward term “witness” (mudu)? And to what extent did Shen’s standard for goodness in formulas depart from that of earlier authors? Because the title of Shen’s formulary is Good Formulas, when he explains in his preface what “good formulas” (liangfang) are, it is difficult to judge at first glance whether he is merely introducing characteristics of formulas that he collected in his formulary or proposing criteria for judging formulas as good in general. Analysis of the structure and content of Shen’s preface to Good Formulas and a comparison of it with earlier formularies shows that Shen treated the elements of “good formulas” that he mentioned in the preface as criteria for formulas in general.

Shen Kuo’s preface to Good Formulas is relatively long in comparison with others that were written before or during the Song dynasty. It opens with a lengthy description of what Shen termed “five difficulties” (wunan) one might encounter in treating disorders. After describing these, he criticizes earlier formularies for neglecting such difficulties and declares that he collected only formulas the effects of which he had “witnessed.” He concludes the preface by explaining his method for listing formulas in Good Formulas.

The five difficulties Shen mentions in his preface are “discerning disorders” (bianji), “treating disorders” (zhiji), “consuming drugs” (yinyao), “prescribing medical formulas” (chufang), and “differentiating drugs” (bieyao or bianyao).¹⁸ Ji, and other Chinese terms (such as bing and hou) that indicated what disorders should be treated with medicine, referred to mental and physical issues that are much broader in scope than the term “disease”—in the biomedical category—implies. Medical authors in imperial China seldom used those terms to distinguish clear-cut biomedical categories of ailments, symptoms, syndromes, and diseases. To indicate the wide scope of reference of these Chinese terms, they are more properly translated as “disorders” than as “diseases.”

Shen Kuo’s explanation for choosing these activities as the five difficulties is so detailed that it occupies almost ninety percent of the preface. The first difficulty involves discerning disorders by examining not only a sick person’s pulse pattern but also his or her complexion and “twelve bodily channels.” The second difficulty involves treating disorders in accordance with a sick individual’s particularities, which might range from age to body size to social status to place of residence and other characteristics, as well as the seasons and the weather when a sick person experienced a disorder, and the progress of that disorder. The third difficulty, “consuming drugs,” pertains to tailoring preparation methods to specific drugs; for instance, some were suitable for cooking for a long time or over a strong fire while others were not. The fourth difficulty involves prescribing medical formulas by attending not only to the specific qualities of each ingredient—to avoid cases in which the effects of a drug might be diminished when combined with other medicinals—but also to a sick patient’s particularities. The final difficulty, “differentiating drugs,” involved discerning between high- and low-quality ones, which was sometimes a function of the geographical features of the places where the drugs originated, the seasons when they were collected, or the methods used to collect and store them.

Concluding his articulation of the difficulties, Shen commented that “the five are merely a sketch. Their intricacy reaches the point that speech cannot express it; their details reach the point that writing cannot record them. How could vulgar persons easily speak of medicine?”¹⁹ Notably, throughout the preface, Shen uses no specific terms to designate the intricate phenomenon toward which the five difficulties gesture—he uses the pronoun “it” (qi) only. His silence further manifests his stress on the ineffable nature of the profound healing arts.²⁰

Concepts resembling Shen’s belief that the subtlety of healing arts defied articulation could be found in the words of the administrator and historian, Sima Guang (1019–1086). He disagreed with the policy of selecting medical officials only via oral tests of their acquaintance with medical classics; he argued that, through the accumulation of experience in treating disorders, “a good physician” (liangyi) could examine his success and failure, investigate the essence of healing arts, obtain the arts via his heart, and learn nothing at all via reading ancient texts.²¹

After narrating the profoundness of the healing arts, Shen Kuo juxtaposed his criticism of earlier formularies with elements of good formulas that he proposed: “Those who compose formulas in the world, when mentioning the therapeutic effects of the formulas, often prefer to overstate it. As for formularies like Thousand in Gold and Kept in One’s Sleeve, they are especially full of exaggeration, which makes people dare not to believe those formularies again.”²²

Immediately after these remarks, Shen wrote the sentence quoted at the beginning of this section, announcing that what he calls good formulas were those the effects of which he had “witnessed.” After this sentence, he shifts back to his criticism of earlier formularies and then introduces another element of his good formulas:

However, people’s disorders are like what I call five difficulties; how could a formula invariably be good? As soon its effects are seen, calling it good; does this method of identifying a good formula bear no difference from marking a boat from which a passenger dropped his sword into a river so as to find the lost sword when he came back to the river again? I thereby detail situations under which a formula had effects at the end of the formula so that one who suffers similar disorders may coincidently match the conditions and implement it.²³

Taking the preceding two quotations together, we can acknowledge the value Shen attributed to case-based knowledge and his disagreement over viewing effects as a single factor for identifying a good formula. The identification pattern over which Shen disagreed is reminiscent of the effective-formula strategy that prevailed in pre-ninth-century formularies. The two formularies Shen criticized were both widely acclaimed medical treatises of the Song period. Kept in One’s Sleeve was The Formulary Kept in One’s Sleeve for Every Emergency (Zhouhou beijifang, or The Formulary Kept in One’s Sleeve to Rescue the Dying, Zhouhou jiuzu, hereafter Formulary Kept in One’s Sleeve). Its author, Ge Hong (283–343), was one of the most prestigious scholars and Daoist practitioners of his day. He completed this formulary sometime between 306 and 333.²⁴ Thousand in Gold was a common abbreviation of Essential Formulas. As early as 932 but sometime prior to 1064, this formulary had been printed, possibly engraved by a commercial publisher.²⁵ The Song court edited Essential Formulas in 1066 and printed this edited version in the same year.²⁶

In addition to mentioning the achievement of therapeutic effects, Shen Kuo proposed two more elements of a good formula. One is an author’s “witnessing” its effects, which concerns the standard for selecting formulas to be collected in a given formulary. Another is providing detailed descriptions of situations where a formula succeeded, which concerns the pattern followed for writing the formula. The two elements functioned as proof that the effects of a given formula are not an author’s fabrication or overstatement, and as instances of specific circumstances in which a formula would have effects.

From the content and structure of Shen Kuo’s preface, we can see that his words regarding the elements of good formulas serve to redefine the criteria for all formulas rather than to introduce features that his book collected. The first indication is the presence of the evaluative rhetoric that he deploys. A comparison of Shen’s preface with Liu Yuxi’s will help us discern the former’s evaluative language. In the latter, the word “trustworthy” that Liu used in his formulary’s title established that the sources of information from which he had learned of the effects of his formulas were well grounded. Throughout Liu’s preface, he avoids commenting on the trustworthiness of other formularies. Like Liu, Wang Gun, in his preface to Broad Relief, merely sounds a note of caution concerning the sources of information that confirm formulas’ effects without commenting on other formularies’ sources. In contrast, Shen Kuo not only compared his proposed elements of good formulas with those in two other famous formularies, but he also evaluated and criticized the two. In so doing, in my reading, Shen featured his proposed elements not merely as characteristics of his own formulas but also as alternative standards for those that deserve high praise.

The structure of Shen’s preface further indicates that he regarded his proposed elements as criteria for good formulas. The preface starts with his articulation of the five difficulties that one might encounter when trying to achieve therapeutic results. Only after spending several paragraphs expounding on these difficulties did Shen then launch into his critique that the authors of earlier formularies employed effects as the single standard of good formulas and propose his own elements. The question “How could a formula inevitably be good?” at the end of the aforementioned quotation nicely encapsulates the gist of Shen’s entire preface, that is, the contingent nature of formulas’ effects in the realm of medicine.²⁷

Shen’s lengthy description of the five difficulties, through the structure of the preface, illustrates the contingent nature of the effects of formulas, hence helping his readers understand the defects in the conventional criteria for a good formula and appreciate why his offer superior alternatives. To better illustrate the structure of Shen’s preface, we must compare contexts where the contingent nature of formulas’ effects is mentioned in the preface and where it appears in Essential Formulas.

The principle that effective treatment required taking all the contingent factors associated with disorders into account was an entrenched idea that existed well before Shen’s Good Formulas. In the early seventh century, for example, Sun Simiao had already remarked in the very first chapter (juan) of his Essential Formulas on the principle that prescriptions should be adapted to patients’ particularities. That chapter comprises nine sections. The names and sequences of the nine sections show that Sun discussed this principle in the context of explaining how to learn and practice medicine; in sequence, they are: “1. Great Physicians’ Learning Practices,” “2. Great Physicians’ Diligence and Honesty,” “3. Brief Examples of Treating Disorders,” “4. Diagnosing,” “5. Prescribing,” “6. Applying Drugs,” “7. Combining Compounds,” “8. Consuming Drugs,” and “9. Storing Drugs.”²⁸

Sun introduces, in the second section, difficulties encountered when attempting to discern disorders and, in the third, proposes a general strategy according to which a prescription should be modified in accordance with the geographic particularities of places where a sick individual had stayed. Sun’s introduction to such difficulties and the corresponding strategy, within the framework of the nine sections, serves in part as a guide to practicing medicine. By contrast, although the five difficulties in the preface to Good Formulas concern a similar prescription principle, immediately following this exposition, we encounter Shen’s proposal that detailing the circumstances in which a formula has succeeded is one of two elements of a good formula. The sequence of Shen’s narrative clearly manifests the context in which he talked about the prescription principle as the proper template for documenting a formula rather than as a general rule of healing practices.

Shen Kuo thus proposed three elements of a good formula: indicating a remedy’s tested effects, documenting an author’s “witnessing” of its effects, and narrating situations where it succeeded. The three are criteria for qualifying formulas. The first was already commonplace in formularies completed before Shen’s day; the other two are relatively new. Shen’s advocacy of the criteria thereby places more empirical weight on the word yan (“effective” or “effects”) in formula writing than it had borne before.

LIMITS OF EXPERIENTIAL VOCABULARIES

Good Formulas comes down to us in an altered form. It was added to a collection of Shen’s other writings concerning medical affairs and formulas, and to prose writings about medicine that the leading literary figure of Shen’s day, Su Shi (1037–1101), composed. The expanded version is known as Good Formulas of the Academicians Mr. Su and Mr. Shen (Su Shen neihan liangfang, hereafter Su’s and Shen’s Formulas).²⁹ The earliest indication of the existence of Su’s and Shen’s Formulas comes from a privately compiled catalog of books, Memoirs of Reading in the Jun Studio (Junzhai dushu zhi, ca. 1151–87, printed in 1180–84).³⁰ Over the course of the twelfth century, Su’s and Shen’s Formulas superseded Good Formulas, becoming the predominant version. The declining circulation of Shen’s original version, as historians have noted, is evident in changes in introductions to this formulary in the other two privately composed catalogs in the Song era. One is The Suichu Hall Catalog (Suichu tang shumu), which was completed by You Mao (1127–1194). Another is Zhizhai’s Annotated Catalog (Zhizhai shulu jieti), which was written by Chen Zhensun (1179–1262). Both catalogs introduce Su’s and Shen’s Formulas without mentioning Good Formulas.³¹

The received version of Su’s and Shen’s Formulas consists of 252 items, 172 of which historians have identified as left by Shen Kuo.³² Among those, 143 are accounts of therapies. Historians generally agree that the 143 items were drawn from Shen’s original Good Formulas. The contents of some of the 143, however, differ markedly from Shen’s criteria for writing a good formula. In the preface, Shen stresses having a “witness” to the effects of a formula as a criterion for selecting it for Good Formulas. In the main text, nonetheless, relatives or acquaintances of Shen attested to the effects of many formulas.³³

Sources that attest to the effects of formulas in the main text, some of which Shen himself notes explicitly and some of which I infer, can be classified into eight types and sequenced on the basis of how closely each resembles Shen’s firsthand experience. Of the 143 items, 30 do not offer information sources for the effects of the remedies in question. Many offer multiple types of information sources.

The first type of source is Shen himself, after applying a remedy to himself or to others and achieving therapeutic effects.³⁴ The second type is when Shen documents how he “personally saw” (in Shen’s words, qinjian or mujian) the effects of a remedy.³⁵ These first two types correspond to our modern notion of having a witness, but in total these types account for only 24 of 143 items. The third type of source is when Shen gave a remedy to someone and that person found it effective (13 items).³⁶ Shen’s own relatives represent a fourth type, constituting the second largest portion of the 143 items (28).³⁷ The fifth type derives from people whom Shen documents in Good Formulas as friends or at least people to whom he had spoken personally (11 items).³⁸ The sixth type is listing the names of persons who had been treated effectively but with whom Shen does not indicate having any particular relationship, which forms the largest portion of the cases (52 items).³⁹ In the majority of the cases that I classify as of the third, fourth, fifth, and sixth types, it is difficult to know for certain whether Shen learned of the effects of their treatments by witnessing them himself or by hearing from those individuals after the fact. The seventh type of source involves secondhand knowledge, that is, hearing from third parties that particular formulas had treated patients effectively (1 item).⁴⁰ Finally, in some cases Shen learned of the effects of a formula by reading about them (7 items).⁴¹

Let us consider several explanations for the inconsistency shown in the abovementioned statistics. One is that Su’s and Shen’s Formulas encompassed records that Shen had not collected in Good Formulas. If we were to count only formulas belonging to the first and second types as those drawn from Good Formulas, the number would be twenty-four. This number, however, is remarkably lower than the current scholarly consensus suggests, according to which most of Shen’s formulas in Su’s and Shen’s Formulas came from Good Formulas. This apparent conflict with the current scholarly consensus reduces the likelihood that this explanation is true.

Another factor that might explain the inconsistencies among the sources of information that Shen claimed in the preface and what is reflected in the main text is that, because in the former he already mentions witnessing the effects of formulas that he had collected in Good Formulas, he did not believe it was necessary in the main text of the formulary to account for scenarios in which he had witnessed a given therapy succeed. If that had been the case, the various types of sources that Shen noted in individual formulas would turn out to have been nothing but his arbitrary choices.

A more grounded explanation of the inconsistency is that Shen saw no fundamental differences in evidential power between validating a formula through personal experience and doing so by reference to others whose information sources he deemed reliable. This explanation can be supported by the typical narrative focus in Shen’s formulas: the historical factuality of the case histories. One example comes from a formula for the Costus Root Pill (Muxiang Wan). In the formula, Shen offers two case histories in which the pill succeeded.⁴² The first involved narrating a healing event in which his elder brother, Shen Pi, was the patient. The second depicted a case that involved Shen Kuo himself.

Shen Pi led soldiers in pursuit of bandits in Zhangpu (around present-day Zhangzhou city in Fujian), and all of the soldiers were suffering from “intermittent fever” (nüe); the Costus Root Pill cured them. Shen Pi was presumably in Zhangpu in 1077, when he was director-in-chief of the Fujian circuit.⁴³ In the same year, Shen Kuo was first provisional commissioner of the State Finance Commission and was then assigned to Xun Prefecture (in Anhui).⁴⁴ He by no means was able to witness the effects of the pill in Fujian. Shen Kuo said nothing about why he narrated his brother’s healing event before narrating his own. This silence denotes that the narrative focuses not on differences between firsthand and secondhand experiences but rather on the historical fact that the pill achieved its effects in specific circumstances.

A relatively indirect piece of evidence indicating that Shen Kuo was willing to accept secondhand testimony as readily as his own experience of the effects of the formulas he included in Good Formulas is that the equivalence between seeing and hearing was widespread across genres in Song China. This equivalence is exemplified in pharmacological texts and notebooks. As I observe in the next chapter in light of more specific evidence, the authors of notebooks and pharmacological writings in the Song era would use what they saw or heard from others whom they deemed reliable to expand on or challenge existing pharmacological reports. One example of this practice can be found in Shen’s notebook Brush Talks from Dream Brook (Mengxi bitan, hereafter Brush Talks).⁴⁵ When accounting for what they saw and heard, those authors rarely prioritized the former over the latter. A shorthand phrase that those authors frequently employed to refer to such information sources is “seeing and hearing” (jianwen, or conversely, “hearing and seeing” [wenjian]). Granted, the phrase “seeing and hearing” expresses equal appreciation of those two knowledge sources far more explicitly than the phrase “witness” (mudu), which Shen employs in his preface.⁴⁶ Despite the difference between the phrases, however, it is difficult to ignore the fact that Shen demonstrates the same equation between seeing and hearing in his contributions to Su’s and Shen’s Formulas.

The main texts of Shen’s formulas that Su’s and Shen’s Formulas collected show that his “witnesses” actually encompass experiences of both the author and others. When presenting formulas, Shen draws no clear line between the evidential power of his experience and that of others. Moreover, in Shen’s description of his or others’ experience, the narratives do not focus on differences between firsthand and secondhand experience. Instead, he emphasizes historical factuality, specifying circumstances in which a given remedy had proven effective. This equation of the evidential power of authors’ seeing and hearing also appeared frequently in notebooks and pharmacological writings during the Song dynasty.

SHEN KUO’S EPISTEMOLOGICAL PROPENSITY

From its preface to its presentation of arrays of formulas, Good Formulas exhibits the same epistemological predilections as Shen Kuo’s other famous work, Brush Talks, which he finished sometime between 1088 and 1095, a period overlapping his composition of Good Formulas. Brush Talks contains records of a wide range of subject matter, including administrative issues, music, calligraphy, painting, natural phenomena, medicine, and so forth. Many of the records fall into domains of knowledge that we now classify under the rubric of science, such as chemistry and astronomy.⁴⁷ The records concerning his discoveries of scientific knowledge place Shen ahead of his time and render him a well-known polymath in modern China.⁴⁸

Although he was a keen explorer of multiple domains of knowledge, Shen strongly doubted the human capacity for fathoming and articulating the ultimate principle of the universe. In Brush Talks, the doubt can be seen most clearly in an entry in which he argues that, in terms of calendric calculations, a human designer cannot precisely predict the “degree” (du) of movement of celestial bodies. Shen adds that “people in the world who speak about such a number capture only its coarse vestiges. However, there is an especially intricate number that cannot be known via calendars.”⁴⁹

Notably, as in the case in which Shen used only “it” (qi) in reference to the inarticulable profoundness of the healing arts underlying the five difficulties, in this quotation Shen again fails to employ specific terms to denote the intricate number. His opinion here of the limited predictive capacity of human-designed numbers to calculate calendrical dates resembles his belief in the limited capacity of vocabularies to transmit medical knowledge. Considering that humans were unable to expound on the profoundness of the universe, Shen passionately observed and recorded particular “things” (wu, including objects and affairs) that he believed derived from the profoundness of the universe. In Good Formulas, Shen elaborated on the contingent nature of a remedy’s effects and detailed the conditions in which a formula succeeded, but made no attempt to address prescription strategies related to the profoundness of the healing arts. In Brush Talks, for example, he recalled how he anticipated the pattern of “circulated qi” (yunqi) by observing changes in the natural environment, such as the strength of the wind, whether it was rainy or sunny, and whether the grass and trees were flourishing or withering.⁵⁰ Shen’s interest in observing and recording particular things is also evident in various projects that he worked on during his bureaucratic career in the 1070s, such as supervising hydraulics, reforming calendars on the basis of his astronomical knowledge, and surveying the Song state’s border with its northern Khitan neighbor, the Liao state (907–1125).⁵¹

When composing Good Formulas and Brush Talks, Shen tended to present his records of particular things in a nonsystematic manner. In Brush Talks, he arranged the records, which covered a broad range of subject matter, in an item-to-item layout without an obviously systematic scheme for organizing them. This manner of presenting knowledge in Brush Talks is understandable insofar as most notebooks adopted a similar approach. The two features together made notebooks an ideal written platform for authors interested in recording various findings that were not systematic or rich enough to constitute an independent book. In comparison, the nonsystematic presentation of knowledge in Good Formulas is especially noteworthy. Among the historical sources I have found, Shen is the first author who explicitly admitted to listing formulas in his formulary in no particular sequence. He says in the preface to Good Formulas that he arranged “sections out of sequence; as long as I was able to obtain a formula, I recorded it and gave it to others.”⁵²

In formularies completed before and during Shen’s time, the conventional organization of formulas involved grouping them by the types of disorders that a given one could treat. The search logic behind this organizational approach was that a reader would first find a section corresponding to specific disorders or symptoms from which people were suffering and then search for an applicable formula in that section.⁵³ By diverging from such an organizational scheme, the nonsequential listing that Shen deliberately employs in Good Formulas was far more challenging for a reader who was eager to find a formula quickly. Shen scribed this unusual arrangement to his eagerness to share effective formulas with the sick, which he claimed left him no time to classify them systematically. Here, Shen’s explanation turns the nonsequential presentation of formulas into a manifestation of his benevolence.

Shen’s epistemological propensity determined the contents and presentation formats of both Brush Talks and Good Formulas. His epistemological approach included the denial of the human capacity to fathom and articulate the profoundness of the universe, an intense interest in observing and recording particular things, and a preference for presenting knowledge in a nonsystematic manner. This trifold propensity, especially in regard to the first two, helps to explain Shen’s emphasis on the contingent nature of the therapeutic effects of formulas and on documenting situations in which a formula proved effective. In addition to his epistemological propensity, other external social and intellectual contexts of Shen’s day likewise fostered the emergence of his new criteria for good formulas.

SONG MEDICAL GOVERNANCE

Before losing the northern half of its territory to the Jurchen Jin state (1115–1234) in 1127 and rebuilding its regime in the south, the Song state, which is known as the Northern Song (960–1127), engaged itself with medicine, education, and public welfare more actively than any imperial government before or after did.⁵⁴ The Northern Song government extended the Medical Academy (Yixue) from its capital to the prefectural level and produced a massive number of medical treatises. It frequently disseminated those treatises and drugs to laypersons across its territory as a form of state relief during epidemics and as an innovative means of educating common people. It also launched many campaigns against the perceived unorthodoxy of healing customs in the south, such as a preference for shamans over medicine. During this period, providing drugs and medical texts to common people was beginning to become a central means of demonstrating Song rulers’ benevolence, exhibiting the state’s authority, and legitimizing its rule.⁵⁵ Provision of medical care became more politically meaningful than ever before.

Modern historians have studied this conspicuous expansion of state involvement in medicine in the Song era, which they characterize as medical governance or medical activism.⁵⁶ Some policies associated with medical governance were integral to a series of reforms designed to resolve social and bureaucratic problems that the Northern Song court undertook, which include the short-lived Qingli Reforms (1043–45) and the influential New Policies (1069–85). From the execution of the New Policies through the fall of the Northern Song in 1127, proponents and opponents of the policies conflicted with each other, spawning venomously factionalized politics. Medical activism nonetheless outlived those factional conflicts. Historians have observed multiple motives that encouraged the long-lived regime of medical governance in the Song era. One was the Song emperors’ personal interest in medicine, especially in collecting medical formulas.⁵⁷ The other motives included the desire to earn high financial profits from selling drugs to imperial subjects, political and medical demands associated with rampant outbreaks and epidemics, and the military’s need to maintain the well-being of soldiers stationed on the borderlands. The last motive was an absolute necessity, considering that the Song state existed under constant military threat from its powerful neighbors.

Few policies that historians regard as integral to Song medical governance exerted influence as lasting as the government’s medical publishing projects. Their profound influence derived mainly from their qualitative and quantitative outcomes. Over the course of 350 years before the Song era, the imperial court commissioned only five medical treatises, whereas, in its 160-year reign, the Northern Song court cut woodblock versions of at least twenty-five medical texts.⁵⁸ The court’s editions of those texts often became the oldest that we have now, and many of the twenty-five texts are still used in today’s traditional Chinese medicine (TCM) curricula in China.

The Song government’s devotion to distributing medical writings marks the first time in Chinese history that rulers attempted to “educate and transform” (jiaohua) the common people by disseminating medical knowledge, especially in textual form.⁵⁹ This devotion arose at the very beginning of the Northern Song dynasty, when the court was still waging bloody campaigns to reunite China. In 973, the founder of the dynasty, Emperor Taizu (r. 960–976), initiated a textual project that involved editing pharmacological texts and compiling them into the encyclopedic Newly Established Materia Medica in the Kaibao Regime (Kaibao xinxiangding bencao). In 974, Taizu asked the same editing team to revise this text, in a year that overlapped with his 974–76 campaign against the Southern Tang state (937–976), whose capital was located in present-day Nanjing (in Jiangsu). The court claimed that it converted the two pharmacological texts into woodblock-printed editions soon after their completion.

In 978, approximately two years after the conquest of the Southern Tang, Taizu’s younger brother and the succeeding emperor, Taizong (r. 976–997), began a textual project that involved compiling an encyclopedic formulary. This formulary, finished in 992, bore the title Imperial Grace Formulary of the Great Peace and Prosperous State Era (Taiping shenghui fang, hereafter Imperial Grace Formulary) and collected as many as ten thousand medical formulas. In 981, Taizong issued an order to compile another encyclopedic formulary.⁶⁰ Completed in 986 or 987 and titled The Formulary for Magnificent Healing and Universal Relief (Shenyi pujiu fang), that work survived at least until the end of the Northern Song dynasty but is now lost. It was impressive in terms of sheer volume, comprising as many as a thousand chapters.⁶¹

Roughly overlapping with the compiling and printing of these two formularies were numerous battles fought against the Liao, the biggest northern threat to the Song at that time. The court carved the two grand formularies into woodblocks immediately after their compilation. The Imperial Grace Formulary became one of the most often-cited in the Song era.⁶² After 1005, when the Song campaigns reached a peaceful conclusion with the Chanyuan Covenant with the Liao, thereby securing its rule over China, the court’s projects that involved compiling and editing medical literature reached a peak during the mid-eleventh century.⁶³

In addition to the court, local government officials were also enthusiastically producing and publishing medical literature. Their discourses on the advantages of distributing medical texts resembles those the court issued. Both often appealed to the political implications of distributing medical texts as a tool for educating the general population as well as to the ethical value of such distribution as a manifestation of benevolence.⁶⁴ The media that the court and other officials used to publish medical texts ranged from scribbled forms to stone inscriptions to woodblock printing.

Changes in the use of printing technology in the late eleventh century further increased the accessibility of imperially commissioned medical texts and expanded their readership. In the eleventh century, court-sponsored medical publications were accessible predominantly to imperial physicians, students receiving imperial medical education, and scholar-officials whose positions placed them in the capital. The Directorate of Education (Guozi Jian, literally, the Directorate of Sons of the State) was the chief institution in charge of printing court-compiled texts. Before 1088, the directorate’s medical imprints were printed in large characters. These required a greater investment in paper and ink, resulting in prices that put them out of reach for lay physicians. In 1088, noticing the restricted availability of these imprints “in the public” (minjian), the Song court began to print medical literature in small characters to reduce prices. These helped increase the availability of Song court–compiled medical literature.⁶⁵ From the second half of the eleventh century onward, imperially printed books spread steadily in greater quantities and at lower prices.⁶⁶

THE RISE OF LAY MEDICAL READERS

The Song court’s and other officials’ devotion to distributing medical treatises facilitated and extended the public transmission of medical knowledge. This eventually enabled the number of self-taught medical learners to reach a critical mass for the first time.⁶⁷ Public availability favored self-taught learners, who collected medical texts and then edited and compiled them into “new” medical treatises. Indeed, these became an important group of medical writers in the Song era. The presence of autodidacts as medical authors indicates that no sharp boundary between “professional” writers and “amateur” readers existed in that period. Readership for medical literature was extended to include people whom authors had not known or contacted in person. Knowledge was transmitted to such readers primarily in written formats rather than, as had traditionally occurred, through physicians’ families and in master-disciple lineages, where oral instruction and practical demonstration played vital roles. In comparison, then, with traditional disciples and those who learned medical care from older members of their families—all of whom could learn through written texts as well as personal contact with experts—the expanding group of self-taught learners depended largely on written words.

Of these, many were lay medical readers, literate people who had learned about medicine primarily by reading texts without the assistance of mentors and often found it difficult to design treatments without consulting those texts. Such readers already existed before the Song dynasty, but they gained greater visibility in Song writings. For instance, in the preface to Good Formulas, Shen Kuo says that he hoped his detailed descriptions of circumstances in which formulas succeeded could enable one who was suffering from a similar disorder to map the condition and apply the formula with good results. Shen’s words sketched out a type of medical reader who was able to read medical literature but unable to design treatments without consulting it.

The same type of reader appears in almost the same context in Wang Gun’s preface to Broad Relief. Before each section of his work, Wang detailed symptoms, a patient’s “pulse pattern and breath” (maixi), or the origin of a particular disorder. In so doing, Wang envisioned “one who did not understand medicine, who can meticulously read the detailed accounts and then adapt it to making therapies.” To be sure, it is plausible to assume that a formulary author might exaggerate the population of lay readers to enhance the authoritativeness of his knowledge or to justify the publication of his new book. This possibility nevertheless does not preclude the existence of such readers.

In addition to mentioning lay readers frequently in prefaces, medical formulas, and postscripts to medical treatises, authors in the Northern Song era began depicting in more specific and vivid terms the circumstances in which such readers consulted medical texts. One case in point comes from the pediatric formula we encountered at the beginning of this chapter. In the formula, Shen Kuo notes how he or his family members tried to find a remedy that would be effective for the sick child by administering formulas collected in The Imperial Grace Formulary one by one. Such a trial-and-error approach was a staple of lay medical readers. The act of consulting The Imperial Grace Formulary moreover reveals their use of treatises commissioned by the Song court.

Other examples concerning the expanding presence of lay medical readers appear in diagnostic instructions that Shen offered in Good Formulas. Those rarely required special skills and could be comprehended and carried out easily. For instance, when recommending the use of a Minor Construct-the-Middle Decoction (Xiao Jianzhong Tang) to treat “depletion cold” (xuhan) and pain in one’s abdomen, Shen instructed his readers to identify the depletion cold simply by pressing a patient’s abdomen: if the patient felt pain immediately but less when stronger pressure was applied, this sensory perception would support a diagnosis of depletion cold in the abdomen.

Following the foregoing instructions, Shen quoted sentences that he attributed to Zhang Ji’s Treatise on Cold Damage Disorders (Shanghan lun, ca. 202–17, hereafter Treatise on Cold Damage) to specify another method for identifying conditions under which the Minor Construct-the-Middle Decoction could be applied. That required that readers be familiar with pulse diagnosis, an advanced diagnostic skill that differed entirely from Shen’s easy-to-apply method. The quotation reads: “When the yang pulse pattern is ‘rough’ (se) and the yin pulse pattern is ‘string-like’ (xian), the method of treating acute pain in one’s abdomen is to first administer this decoction.”

Here we see technical terms associated with pulse diagnosis, which was an important technique in classical medicine in imperial China and should be distinguished from pulse-taking in present-day Western biomedicine. The latter focuses on the rapidity and regularity of a patient’s pulse, whereas pulse diagnosis in classical Chinese medicine applied to a variety of conditions that involved relatively subjective tactile perceptions on the part of practitioners. When conducting pulse diagnosis, practitioners typically placed their fingers (usually three) on the underside of the patient’s forearm below the wrist. Imperial Chinese medical texts described the tactile perceptions gained from pulse diagnosis using a rich array of verbs and adjectives, such as “sinking” (chen), “floating” (fu), and “deficient” (xu).⁶⁸ It would take years, even decades, for a diligent medical trainee to master the art of pulse diagnosis. In comparison, the procedure and results of the first diagnostic method that Shen noted were so straightforward that any literate layperson could follow them.

The increasing presence of lay medical readers in Song writings indicates to a certain degree the growth of this readership in the Northern Song era. Extant data do not allow us to gauge the specific extent to which the lay medical readership had grown. Nevertheless, reported figures indicating how many candidates were taking civil service examinations provide some sense of the growing number. Examination candidates constituted an important group of people who were literate enough to read medical literature or might be interested in medicine. The frequent regional trips that those candidates took between their hometowns and examination venues created a need to learn about medicine or at least to be able to apply basic self-treatments on the road. Of course, not every examination candidate had such an interest or was able to access medical literature. We do however see that, starting in the eleventh century, some candidates abandoned their fruitless attempts to pass the examinations and chose to earn their living practicing medicine.⁶⁹ It is probably safe to suppose that many of the candidates consulted medical texts.

One physician who had been an examination candidate, Dong Ji (fl. 1102–1117), even wrote the first formulary devoted to travelers in China, The Formulary for [Ones Staying at] Travel Houses, a Provision of What Is Essential (Lüshe beiyao fang). Travelers whom Dong mentioned as the targeted readership of his work included officials who traveled from a current administrative position to the next one, usually every three years, and “guests” (ke) on the road, whom Dong did not specify but might have included merchants and examination candidates.⁷⁰ Dong’s formulary did not survive, but Qing officials reconstructed one chapter (juan) from the Ming court–compiled encyclopedic work, Great Compendium of the Yongle Era (Yongle dadian, 1403–8). Differing from Shen Kuo’s Good Formulas, the lone extant chapter of Dong’s formulary collected no narratives of healing events. But some of his formulas did concern disorders that might have occurred during trips, such as encountering coldness or heat stroke.⁷¹ Dong’s frequent trips as an examination candidate might have inspired him to compose this formulary.

Civil examinations during the Song dynasty, to be very brief, took place every three years on three levels: the prefectural, the departmental (at the capital, for those seeking advanced-scholar degrees), and the palace (for degree holders to be assigned to their bureaucratic posts according to their ranks). In the early eleventh century, there were approximately 20,000 to 30,000 candidates standing for prefectural examinations, but the figure soared to 79,000 a century later and exploded to 400,000 by the middle of the thirteenth century.⁷² When gathering at prefectures for forthcoming examinations, candidates often were more easily able to access medical texts than they would have been in their hometowns. Even if only half of the candidates in any given round of tests could access and read medical texts, the number is still considerable.

In imperial China, a lay medical reader, if he continued to improve his healing skills, could later proclaim himself a physician and practice medicine as the source of his livelihood. During that era, no institutional qualifications or training served as the gatekeeper to determine who could practice as a physician, nor did any controls differentiate qualified practitioners from incompetent ones. Anyone capable of practicing certain healing skills could call themselves “physicians” and set their own fees.⁷³ Becoming a physician became easier with the increasing availability of medical publications from the Song period onward. This ready availability helped increase the number of self-taught medical practitioners in the Song era.⁷⁴ Dong Ji, for example, was an examination candidate and self-taught medical learner before staking his livelihood on medicine and composing The Formulary for Travel Houses. Bearing this in mind, here the phrase “lay medical readers” serves as shorthand to refer to medical learners with limited healing skills, reflecting the possibility that skill levels changed over time.

When Wang Gun and Shen Kuo touted the benefits of their detailed descriptions of situations in which a given formula succeeded, they both mentioned the benefit of helping their readers judge the suitability of the remedies they were recommending. On the one hand, helping such lay readers, who were unable to design suitable treatments without consulting formularies, might contribute to the authors’ justification for writing a new formulary. On the other hand, the extended public transmission of medical literature and a critical and ever-growing mass of examination candidates in the Song era provide a solid foundation for the existence and expansion of an audience of lay medical readers. Starting in the eleventh century, descriptions denoting such readers appeared with greater frequency.⁷⁵

SCHOLAR-OFFICIALS, THE NEW MEDICAL AUTHORITIES

In 1057, the “military affairs commissioner” (shumi shi) Han Qi (1008–1075) submitted a memorandum in which he suggested that Emperor Renzong (r. 1022–1063) order “Confucian officials” (ruchen) who understood medical texts and imperial physicians to work together to examine and distribute an array of medical texts: Divine Pivot (Lingshu), Grand Basis (Taisu), A-B Canon [of Acupuncture] (Jiayi jing), The Divine Farmer’s Classic of Materia Medica (Shennong bencaojing, hereafter The Divine Farmer’s Classic), Essential Formulas, and The Imperial Library Formulary.⁷⁶ Historians have dated the establishment of the Bureau for Editing Medical Texts (Jiaozheng Yishuju) to Han Qi’s memorandum.⁷⁷ Historians have proposed that the bureau was established as rampant epidemics were breaking out, between 1023 and 1063; the parallel timing suggests that the purpose of founding the bureau was to edit medical texts that would be useful in fighting epidemics.⁷⁸ From 1057 to 1069, it compiled one pharmacological encyclopedia and edited ten medical texts. No evidence indicates that the bureau submitted texts after 1069, which suggests the end of the project. The bureau-edited medical texts went beyond serving as textbooks for imperial medical education because the court promptly printed at least nine of the eleven for the purpose of distributing them widely to the public.⁷⁹ The establishment of the bureau represents the peak of the Song court’s investment in distributing medical texts. After it ceased operations, the court never again launched such a large-scale project related to medical publications.

It is noteworthy that the court assigned Confucian officials, rather than medical ones, the task of supervising the editing projects that the bureau conducted, which institutionally recognized scholar-officials’ authority over the editing of medical texts. Before such assignments were made, there was a long-standing practice whereby non-medical officials compiled and annotated medical texts in private hands. At the institutional level, however, when the Western Han court was for the first time in Chinese history (between 26 BCE and 7 BCE) compiling a catalog for the imperial library, which was titled Seven Catalogs (Qilüe), it assigned a “medical attendant” (shiyi), Li Zhuguo (fl. ca. 26 BCE), to edit books concerning “remedies and techniques” (fangji), and the infantry commandant Ren Hong to edit military texts.⁸⁰ These assignments were clearly based on the editors’ expertise in specific domains of knowledge.

In the eleventh century, court officials forcefully asserted that, regardless of the domains of knowledge to which texts that the court was going to edit belonged, only civil officials in “academies and institutes” (guange) should be in charge of editing projects.⁸¹ As shown in relevant memoranda left by court officials at the time, their decision reflected the belief that text editing was not only an administrative duty that officials in academies and institutes should carry out but also a process by which the officials cultivated a broad spectrum of learning.⁸² The bureau’s basic procedure for editing medical texts involved first collecting multiple renditions of a given treatise, then comparing the contents of those renditions, and finally teasing out the best possible words or phrases where discrepancies occurred; this was the same procedure that officials in academies and institutes conducted when editing Confucian classics.⁸³

Scholar-officials’ newly recognized medical authority reflected their capacity to decide which medical writings were useful or significant enough to be edited and then distributed by the court. This authority is evident in the administrative power by virtue of which they could suggest to the emperor which earlier medical treatises were worthy of the state-sponsored editing projects. The Bureau for Editing Medical Texts was founded in accordance with an edict by Emperor Renzong; however, Renzong left us no edict in which he specified which texts the bureau should edit. Instead, many were nominated in officials’ political documents, as exemplified in the abovementioned memorandum issued by Han Qi. This point is made in a preface as well: when its staff finished editing Zhang Ji’s Treatise on Cold Damage, the bureau composed a preface to the edited version. At the end of it is a list of those who were responsible for the editing, all of whom were scholar-officials rather than medical officials. The preface notes that “the state decreed Confucian officials to edit and correct medical texts … We considered that, regarding the urgency required by hundreds of disorders, none is more urgent than cold damage. Now we first edited Zhang Zhi’s Treatise on Cold Damage.”⁸⁴ This sentence indicates that it was Confucian officials in the bureau, rather than medical officials, who bore the responsibility for deciding which medical texts would be edited.

Only after understanding the medical authority granted to Confucian officials through the establishment of the bureau can we understand the weight of the prefaces they wrote for the bureau-edited Essential Formulas. This version, which was completed in 1066, includes both a preface and a “post-preface” (houxu). The two end by listing the names of Confucian officials in the bureau, the first three of whom are Lin Yi, Gao Baoheng, and Sun Qi. The earlier draft of the two prefaces was written by another Confucian official, Su Song. Historians have compared the two prefaces with those of Su Song thoroughly, and my discussion here focuses on how they praise Essential Formulas.

The four prefaces all cited the advantage of the breadth of textual references that Sun Simiao included in his Essential Formulas. For instance, Su Song complimented the formulary for its “profoundness and broad coverage.”⁸⁵ Lin Yi and other officials also applauded it for encompassing references “drawn from as early as the invention of writing to as late as the Sui dynasty. It did not fail to collect either any classics or formulas. It amassed what numerous healing currents secretly valued.” Another advantage that Lin Yi’s preface noted is that, regarding its formulas, “none of them cannot perfectly verify their effects.”⁸⁶ Su Song additionally appreciated that fewer than one in ten medical texts that were completed between the Han and Tang periods survive to his day; only Essential Formulas, “whose beginning and end roughly appear and is especially a complete book,” remained available.⁸⁷

The three advantages of Essential Formulas that the four prefaces praised were the broad coverage of its references, the effects of its formulas, and its longevity (having remained intact for a long time). The first two also appear in the bureau’s prefaces to other medical texts that it edited. For instance, in the preface to the edited Treatise on Cold Damage, officials cite the historian Huangfu Mi (215–282), who praised it because the application of its formulas “frequently had effects.”⁸⁸ In the preface to the edited The Imperial Library Formulary, officials praised Wang Tao for “collecting the important parts of formulas belonging to numerous currents,” a judgment that agrees with the first recognized advantage of Essential Formulas.⁸⁹

Interestingly, the above officials’ praise of Essential Formulas contrasts sharply with Shen Kuo’s criteria for good formulas.⁹⁰ The two had divergent opinions regarding the appropriateness of applying phrases that remark on the universal effects of a formula as praise. Shen clearly opposed this application. Officials in the bureau, while they accepted the contingent nature of a formula’s effects, felt more comfortable than Shen with employing such phrases. Shen moreover differed from officials in the bureau in terms of the scope of the references or formulas that a formulary contained. If an author collected formulas whose effects had been proven only by himself or by his acquaintances, the number of formulas and types of disorders that his treatise covered was inevitably limited. The limited number of formulas and the correspondingly limited coverage of disorders is reflected in Shen’s Good Formulas.

No textual evidence suggests that Shen Kuo had read the four prefaces to Essential Formulas. On the basis of Shen’s bureaucratic career, however, I suspect that he at least knew that the bureau edited Essential Formulas; his criticism of this treatise therefore could be regarded as an implicit challenge to other Confucian officials’ standards for selecting medical texts. The period during which Shen was an official in academies and institutes overlapped with part of the period during which the bureau fully functioned. In 1066, when the bureau submitted Essential Formulas to the emperor, Shen was the “editorial assistant for books at the Institute for the Glorification of Literature” (bianjiao Zhaowen Guan shuji). He was appointed the “proofreader in academies and institutes” (guange jiaokan) in 1068. Within a month of that, though, he left the position following his mother’s death. Insofar as the period during which Shen was an official in academies and institutes overlapped with the year when the bureau edited Essential Formulas and the court printed it in 1066, it would be surprising to learn that he never knew at that time that the bureau had submitted the formulary to the emperor.⁹¹ Considering that Shen presumably knew that the bureau edited Essential Formulas, his choice of that book as an example that merited criticism may not be simply a response to this treatise’s popularity at the time; rather, it may be also a product of his implicit disagreement over Confucian officials’ standards regarding which formularies should be edited and distributed by the court.⁹²

BEFORE THE NINTH CENTURY, the predominant persuasion strategy used in a formulary was to note that the recorded formulas were known to be effective. Formulary authors who applied this strategy rarely viewed noting who attested to the effects of a formula as central to its being deemed good. Diverging from those earlier authors, Shen Kuo in Good Formulas criticized that strategy and proposed two new criteria: an author’s “witness” to the effects of a formula, and case narratives depicting situations in which the formula succeeded. We should not, however, jump to the conclusion that Shen elevated the role of firsthand observation in writing formularies. An analysis of Good Formulas shows that the “witnessing” that Shen upholds in his preface actually encompassed information sources drawn from both his and his acquaintances’ personal experience. The wide range of information sources that the term “witness” represents indicates that Shen saw no fundamental differences between testimony based on his own experience and that based on the experience of people whose information sources he deemed reliable. Rather than differentiating between firsthand and secondhand testimony, Shen’s “witnessing,” along with case narratives, focused on proving the historical factuality that a formula had demonstrated effects. It is the historical factuality, rather than authors’ expertise in medicine, that underpinned the credibility and reliability of Good Formulas.

Scholars have analyzed the distinction between firsthand observation and secondhand experience extensively in analyzing claims to knowledge in the history of science and medicine in premodern Europe. Their findings reveal a perplexing absence of distinct terminology indicating firsthand observation, an absence reminiscent of what we have encountered in the preceding investigation into Shen Kuo’s phrase “witness” and into medical texts and notebooks of his day. In ancient Greek, the Hippocratic physicians recognized firsthand observation as a source of information.⁹³ Nonetheless, within their corpus, there existed no specialized language that denoted observational practice. What modern scholars have translated as “to observe” does not correspond to any single consistently used vocabulary element. Even as late as the fifteenth century, formulary authors still did not seem to embrace the notion that firsthand experience should be demarcated from secondhand evidence and be clearly labeled accordingly.⁹⁴

Shen’s proposal for new criteria for Good Formulas derived from his epistemological propensity as much as from changes in his surrounding social and intellectual circumstances. He doubted the human capacity to capture the ineffable profoundness of the healing arts. This doubt sparked his keen interest in observing and recording particular things that he thought originated in profoundness, such as specific conditions in which a formula achieved therapeutic effects. Shen attributed his advocacy for describing such conditions to his aim to help readers map them onto a given sick person. This then would enable readers to find a formula that was suitable for that person’s particular bodily condition. Shen’s targeting such readers as the chief audience of Good Formulas paralleled the ever-growing population of lay medical readers in his day.

Imperial recognition of scholar-officials as a group with medical authority helps us understand Shen Kuo’s confidence in and engagement with listing criteria for good formulas. This imperial recognition owed largely to the eleventh-century court’s project of editing and printing formularies, which established the Bureau for Editing Medical Texts and assigned Confucian officials to lead the projects. This assignment institutionally recognized Confucian officials, and by extension all scholar-officials, as a group in which it invested medical authority. Their recognized areas of authority included correcting medical documents, as well as deciding which medical texts were useful and important enough to be disseminated across China by the government.

Although Shen assertively promoted the new criteria for good formulas, their influence on later Song medical writing is no more obscure than that of Brush Talks on later Song notebooks. As scholarship has noted, Shen had already used reliability to assess Brush Talks; nevertheless, no authors of later notebooks that were completed in the Southern Song era identified Shen as a significant contributor to this trend when applying reliability as a standard of assessment in their writing. What historians can best say in terms of the influence of Brush Talks is that Shen “provides a solid precedent in the immediate past that Southern Song literati were eager to invoke.”⁹⁵ The same could be said of the role that Good Formulas played in the development of medical writing: although many later Song medical authors likewise documented witnesses to a remedy’s effects and wrote case narratives to prove them, they rarely attributed this practice to Shen Kuo.

The obscurity of the influence of Good Formulas on Song medical writing by no means impugns the historical significance of that treatise. The treatise signals, in the domain of medical genres, the coming of a new configuration of knowledge production featuring unprecedentedly wide dissemination of medical literature through government and private publications, a soaring population of lay medical readers, and scholar-officials as a group invested with medical authority that for the first time received state support and institutional recognition.

The emergence of this configuration owed largely to Song medical governance, yet its impacts were much more profound than the Song emperors and officials intended.⁹⁶ One product is Good Formulas. This formulary offers us a rich account of how a new pattern, in which authors documented empirical evidence pertaining to a formula’s effects, emerged in response to this configuration. Twenty years after the completion of Good Formulas, a medical author collected his case narratives in his pharmacological work. Its emergence against the new configuration is the subject of the next chapter.