Term Chaos is Just Chaos: Part Four

Submitted by Robert Felt on July 12, 2007 - 4:02pm.
Part Four: The Cost of Chaos

This is the last blog post in the "Term Chaos is Just Chaos" series, the previous posts are:

Part 1: Where the Infringement Hides

Part 2: Not Just a Matter of Words

Part 3: Standards Are Not Shallow

Those who have read the "Guided Tour to the Term Debate" essay on this site, understand that I see term chaos as just the latest step in a twenty-plus year attempt to justify paraphrase and simplification as the primary strategy for educating clinicians in the English-speaking world. The earliest assertion was that Chinese medicine had no terminology, an idea that I feel derived from a lay view of Chinese medical language and a certain embarrassment about Chinese medical ideas that expresses itself in the urge to biomedicalize and de-moralize. Later, as Wiseman and others consistently refuted this notion, anti-consistency arguments migrated, arriving today at "term chaos," which is itself a euphemism for an undocumented plurality that is supposed to be good, despite the lack of any orderly plan for how students are to be taught.

With this post I will bring the "Term Chaos is Just Chaos" postings to an end, some nine months after the conference that inspired the topic. I hope that what I have accomplished with this series is to direct peoples' attention away from slogans, thus bringing to light the losses our profession suffers from chaotic and disrespectful treatment of our field's conceptual foundations in translation.

Counting the Costs

Standardization has costs and they must be considered, if only so that people understand that there's more to the issue. For example, fully documented books with inclusive glossaries and a clear, easily-followed relationship to a Chinese language source are more difficult and expensive to produce. The more depth of a Chinese text you attempt to replicate in English, the greater the chance of error. Thus the time needed, the authorial and editorial skill required, the proofing and fact checking demanded, are more difficult and expensive. It is much easier and requires less knowledge of Chinese medicine to paraphrase a Chinese text than to actually translate it because the more difficult and subtle issues can be personalized or eliminated. In fact, the work is complex enough that it is best done with teams that include both scholars and clinicians. Many more people have the skill to describe their own clinical experience than have the skill to assemble a treatment of a clinical topic where disciplined research is required. Participating in an open standards approach to publication costs more, takes more skill and requires more time. Were the field's authorities to prioritize standards-compliant texts, there would need to be new investments in curriculum and testing. There are indeed costs.

Nonetheless, the lack of standards costs more. Here are a few of the losses our professional already suffers:

● Useful, accurate inter-generational, inter-practitioner case records are impossible because of inconsistent term use and simplification. The "test of time" is lost.

● Software-assisted translation of Chinese databases and clinical records, research and clinical statistics is and will remain underdeveloped, until there is a sufficient level of standardization to inspire someone to invest in its development.

● Clinical trials based on Chinese medical diagnoses and therapies are always suspect and will be broadly dismissed until there are standard terms and definitions sufficient for evidence of inter-rater reliability to be amassed.

● Until standard terms and definitions are in common use, clinical trials based on Chinese medical diagnoses and therapies will be more difficult and expensive to develop because the participating clinicians will need to first master a coordinated clinical language. Trials conducted without such a common language will be unreliable by definition, or based solely on biomedicine

● Translation, even of technical documents, will remain inconsistent, providing less-than-accurate clinical data.

● Databases using non-standardized data, or created from differing translations, will be structurally incapable of high degrees of accuracy, producing false positives ("hits" that aren't), and false negatives ("hits" that should have been) because of inconsistency.

● Relationships to western medical correspondences will not achieve high degrees of accuracy or depth.

● Western medical terms will remain the de facto standard of third-party reimbursement and clinical research.  The longer we go without source-oriented term standards the more the scientization of Chinese medicine will dominate the field.   Organizations like the W.H.O. and S.A.T.C.M. are ahead of the field as regards term standards.  However, their aim is strongly influenced by a priority on publications such as biomedical journals, where fidelity to traditional concepts is less important than correspondences to biomedicine.

● East Asian experts will never consider western work as equal to their own.

● English language clinical and didactic texts will continue to lack information found in their Chinese sources.

● Chinese medical concepts will continue to acquire meanings different from those of the Chinese language definitions.

● The English-language understanding of the relative importance of Chinese medical concepts will continue to differ from the Chinese understanding.

● In clinical practice, students will be unable to recognize when there is more to a topic than they have learned. Without rigorous, unsimplified clinical descriptions it is impossible to know when a concept has been fully explored.

● In-school and license examinations based on undocumented terminologies, simplifications and paraphrases will reduce teacher's options by demanding use of these defacto standards.

In essence, without English language concepts equivalent in sensitivity and quantity to Chinese medical Chinese, English language Chinese medicine will always be less capable, less flexible and less complete. It is like trying to stuff two suitcases of your possessions into a handbag, some things will be broken, some things will be rumpled, and a lot will be left behind.

The Problem of Reliability

Practically, all the losses mentioned above have the same source. If there is no rigorous equivalence between concepts in Chinese and what appears in English, those who depend on English will get an inaccurate and/or incomplete view. But, for the sake of example, let's look at the often-expressed demand for electronically searchable information.

When I was investigating the large scale piracy that took place last year, one of the things I frequently heard from people was that the need for electronically-searchable information justified a degree of looseness with copyright, (some being looser than others). This desire more fairly expresses itself in the frequent demands for electronically searchable text and databases. Although these tools are often seen as being different from classic transmission tools such as books, reports, statistical analyzes and written records, each of these is also essentially a database, a collection of related records stored in natural language.

The most basic step in the creation of a database (or an accurate natural language compilation) is to insure that the design is sufficient for the size (number of records) and scope (extent of each record) of the data to be stored and/or transmitted. The problems we face today are not really a problem of terms but instead the side-effects of early, inadequate analyzes of Chinese medicine's conceptual size and scope (e.g: the idea that Chinese medicine's terminology is limited to unfamiliar terms). Since the first assessments of Chinese medicine greatly underestimated its conceptual richness, the strategies adopted for concept transmission were overly simplistic and inadequate for the actual scope of Chinese medicine. To understand this problem, it helps to think of information about Chinese medicine as an on-line encyclopedia of concepts, medicinals, acupoints, formulas, etc.

The simplest, most reliable strategy for storing information is to give each record a unique "key" - a value that uniquely and always identifies that information. In fact, it is not just the simplest and best way; it is the only way because database descriptive languages (sometimes called "data dictionaries") do not permit multiple records with the same primary key. Because software (including the human software of memory) must be able to find the information needed; that is, to select the proper record from a database, using the same key for multiple records is an absolute mess. This is why there are arbitrary keys like sequential numbers for much of what is stored and/or processed by computer. Everyone has account numbers, customer numbers, and other identifiers like social security, drivers license, and bank IDs. These are unique keys to your various personal records.

While there may be many people with your name where you live, only you will have your driver's license number. A database of Chinese medical ideas or data such as medicinal or formula names is no different. While it is true that records can be accessed by different names, this is really not an exception to the unique key principle. Rather it depends upon accurate recognition of secondary or compound identifiers. Imagine an on-line Chinese medical encyclopedia that people could search using the words found in their books. Now, consider that even for some very basic ideas such as the five minds and seven affects, there are differing terms. 忧 (yōu) is translated by Wiseman as "anxiety," by Chen as "melancholy" and by Maciocia as "worry." The unique key is actually the Chinese (忧) because it is the uniquely-identifying source. Thus a search engine that needed to deal with these multiple English terms for a single concept would need to be programmed to recognize all three terms - anxiety, melancholy, worry - as related to yōu.

We can demonstrate this in pseudo code[1]; for example, one resolution to the problem of these multiple translations might be to add the author to the search word for a compound search:

IF SEARCH-TERM = "anxiety" and SEARCH-AUTHOR = ‘Wiseman' THEN DISPLAY "yōu"

IF SEARCH-TERM = "melancholy" and SEARCH-AUTHOR = "Chen" THEN DISPLAY "yōu"

IF SEARCH-TERM = "worry" and SEARCH-AUTHOR = "Macioca" THEN DISPLAY "yōu"

This is further complicated by the fact that "worry" is also used by Chen for思 and the fact that in paraphrased texts there is no guarantee that every use of 忧 or 思 will be recognized or consistently named. This is a relatively simple term with a generally-agreed definition. The pseudo code should give you a slight idea of what it takes to resolve these inconsistencies in order for a database to function. Clearly, if there are no open lists of equivalents, or the equivalents are inconsistently used, searches will never be accurate. This should also give you some sense of how much more critical inconsistency becomes when thinking about text searches in natural language texts and the storage and retrieval of more complex notions with multiple, source-determined, or historically-various meanings. Ironically, whatever value there may be in plural translations depends on open standards - published lists of equivalents - and is completely erased by paraphrased text.

The critical points are:

1. Where there are no sufficient published lists (an open standard) for a body of literature, there is no reliable way to search that literature. Whether using a written index, a search engine, or a database the information you receive will be no more accurate than the text is consistent. Software can make it easier to search but software cannot fix the inconsistency. This is the root meaning of the programmer's expression "GIGO" (Garbage In, Garbage Out).

2. Where the number of terms (keys) used in a body of literature is too small to cover the number of concepts (records), accurate searches are impossible. Information created through an inadequate assessment of extent and scope, no matter how accessed, will be incomplete and inaccurate.

3. Regardless of any term list, information that is paraphrased rather than translated will be more inconsistent (i.e: have a greater number of false positives and false negatives).

So, who cares? You do, if you believe that Chinese medicine as it has evolved in East Asia is worth knowing. You do, if you believe accurate and reliable electronic data will be of use to you. Yet, there are also other professional issues. If you have looked at some of the research by which our medical abilities are challenged, the reduction of Chinese medical concepts to not-quite-equivalent biomedical concepts has played a significant role in "debunking" Chinese medicine. Chinese medicine could never have evolved without an intergenerational case record, so the idea that undocumented and inconsistent concept identification is some sort of value is, frankly, nonsense. Without an effective inter-generational literature, we can throw the "test of time" in the trash. Without a case record that can be consistently recorded in information systems, readable by any clinician (at least any clinician with a dictionary), and accessed by people trained in different schools and styles, Chinese medicine can not be integrated into the real world of medical delivery. Without an intergenerational shared language, precise communication, and the ability to show reliable outcomes, how do we deal with the reality of medical economics?

Term chaos, just won't get us there. But an open standard can.

In conclusion, I want to thank the people who have made more than a thousand reads of this series. My personal hope for the future is that we will arrive at a Chinese medical education that produces an at-scale profession capable of contributing to the vast health care needs of the world. I have nothing against individuals or schools that want to do something else. Wanting information to be treated in standard and respectful ways is not the same as wanting everyone to do the same thing - despite the steady "spin" that it is. Throughout history native Chinese have acquired a depth of clinical skill with educations that have consisted of rote memorization, apprenticeships, as well as within the strictures of a modern system that is biomedically dominated. This is possible because Chinese speakers have always had a standard -- Chinese medical Chinese. Until we come to recognize the value of that standard, and to abandon clinging to early and under-estimated assessments of its size and scope, we will fail to realize Chinese medicine's full potential.

[1] People familiar with coding will know that this problem would likely be solved using cross-reference tables rather than the example pseudo code. The example merely shows how unstandardized terms or terms not referenced to the Chinese source create inaccuracies. It is not a practical problem solution but an example of the problem itself.