TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES

Here is the original article from UBS TTW 1984:

Supplemental material to the article:

Anne Kompaore has added value to this post by overseeing the transcription of Harold Fehderau's original supporting material to his 1984 TTW article above and making it available. Harold's project cannot truly be appreciated without seeing the full extent of his data. Whatever you might think of the idea of this kind of micro-checking, it is an impressive piece of work.

Here is Anne on the subject in her own words (from a recent email):

"I think both consultants and translators need a kind of checklist, at least near the beginning of their work, in order to keep in mind the things to look for. A good team should need micro checking (not micro management) only in the first stages of their work as they learn the ropes. Consultants need to be aware of all the possible issues and check for them initially at some point. If he/she is satisfied that the team is managing certain aspects, then he does not need to check for them anymore. Unfortunately, though, there are too many teams out there who are not doing a good job for one reason or another.__"

Extracted text from TRQUAL.doc:

UNITED BIBLE SOCIETIES TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART

For discussion only Not to be published in this form.

TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES (Reference Manual for Bible Translation Checking) AUTHOR: Dr Harold W Fehderau When a consultant is called upon to examine a translation for publication, his evaluation should aim to be exhaustive, including aspects such as the textual, exegetical, semantic, and grammatical accuracy of the translation, the contextual consistency of key words, the adequacy of its style and discourse structure, and the acceptability of the orthography and format of the manuscript. A new consultant usually begins his work by making a number of trips with a senior consultant to observe how the work is done--and then he is on his own. His work can be quite frustrating when he finds out that there is no systematic guide to verses and passages in the Bible that will give him a significant insight into a translation for checking purposes. He sees that consultants with more experience have a number of favorite passages that they turn to when asked to give an expert opinion on a translation, especially when there is only a limited time for checking. (Many of these passages happen to be the first chapter of Biblical books: Gen 1, Ps 1, Mk 1, Jn 1, Rm 1, Gal 1, Eph 1--all of which present a wealth of translation problems.) However there is no place that a consultant can turn to for systematic help. There is no way of knowing which passages are particularly comprehensive for checking, and after a checking session, there is no way of assessing whether a complete picture has been gained of how all the possible problem types have been handled by the translators. Furthermore, we find that even experienced translation consultants are often very good at checking over some parts of Scripture (particularly the Gospels) but they feel less sure of themselves in other parts, particularly in the the Old Testament. What is needed is a systematic guide to the most significant diagnostic passages for checking each particular book of the Bible, and a set of passages to check entire Testaments in an efficient way. The result should be an accurate picture of the quality of a translation in all its aspects. We need help in measuring how close to the bull's eye a translator has come in aiming at the target language. Even after working on checking a translation for two or three days, a consultant sometimes finds that the translators have not fully understood the impact of the sessions, even when the translation is really quite inadequate. There are various reasons for this, but one of them could be that the consultant had no really systematic and convincing way of pointing out the problems and needed solutions.

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES

PAGE 2

A reference manual offering help in this area would systematize (if not standardize) checking procedures. It would be particularly helpful to new consultants at the very beginning of their work when it is most needed. And with proper instruction, it could also be used by experienced translation teams to evaluate their own work to see if high quality standards are being met. It often happens that more than one translation consultant is called upon to work with a particular language project (over the course of time). Consultant reports on the checking sessions are usually given in only very general terms, and sometimes differences in approach and thoroughness of investigation make it quite hard for the next person to assess what exactly has been done and what still remains to be done in the checking work. In other words, we have no systematic, standard procedures for evaluating a translation, and there is no guarantee that a translation approved by X would also have received the approval of Y had he done the checking. We need a guide that will ensure thoroughness of our checking procedures and that will yield an evaluative report that can be understood and endorsed by any other consultant. QUALITY CONTROL AND QUALITY ASSURANCE r Once translators have been trained and have begun their work, a consultant should have frequent contact with the team. Consultant checking sessions should not be considered a visit from an "inspector", but rather an opportunity to give additional help to do high quality work. Checking sessions should be designed not only to check out the accuracy of a given number of items, but also to reveal the areas of strength and weakness of the team. In other words, checking sessions should aim at teaching the team to do better work. Team members must learn to apply translation principles independently. They should be able to identify problems and come up with solutions. They should be able to apply linguistic analysis techniques and make their way through difficult passages. They need to be able to control the quality of their work as they translate. The emphasis in a checking session is not so much on, What did you do wrong? but on, How can you do even better? Translators look forward to a consultant's visit when the accent is on training rather than inspection. Translation checking is particularly crucial during first draft work. Revision work is of course much easier when the first draft is good. Too many committees rely on revisers to improve their work and catch the mistakes. Translation checking with an eye to on-going training it is problem-oriented in a systematic way. There are problem areas in Bible translation; checking sessions of these has been handled so that additional training items that have not been mastered. is most meaningful when a number of known should reveal how each can be given on the

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES

PAGE 3

Steps are being taken to provide a guide for checking (in the form of a Reference Manual) that defines translation problems and lists scores of passages for each problem type in the various books of the Bible. (See the attached rough draft sample list attached: "SAMPLE LIST OF TRANSLATION PROBLEMS". Periodic random checks will show whether the team is improving in their handling of each particular problem. This approach has the advantage of concentrating on particular items and letting the cumulative effect of many examples indicate the degree of mastery and understanding of the problem and its solution on the part of the translators. It is also necessary, of course, to check connected passages and long stretches of material in order to evaluate certain aspects of a translation that cannot be dealt with in the problem-topic approach. Ideally every verse should be checked by a consultant. This has not usually been possible for UBS personnel because of the limited time that can be spent with each project. This means that checking has to be done by random sample methods, spotchecking here and there. Spot-checking done in a systematic way can cover all of the problem types encountered in translation. When a particular problem type is done well consistently in a large random sample, it can be assumed that the cases not checked have also been done well. (This is a basic assumption of this approach.) WHEN IS A TRANSLATION ACCEPTABLE? We should have procedures that are standard enough that any consultant will come to essentially the same conclusion about the quality of a particular translation. My assessment is that this is not the case at the present time. Matters of language level and acceptability of usage and style cannnot be evaluated by a consultant in the same way as other aspects of a translation. For these he must rely on the evaluation and approval of native speakers of the language based on good testing methods. (See also the section below on "Testing Translations" in the Supplementary Material.)

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES

PAGE 4

Assuming that those matters have been taken care of, it is my experience that a translation is acceptable and has reached a high quality of work when a random check of hundreds of examples of all possible translation problem types yields a score of 95% or more items done correctly or at least acceptably. Note that translations that give a literal rendering of a King James or RSV base text, do not stand much of a chance of getting a score of over 40% (which means that out of a sample of 100 varied problems checked by a consultant, only 40 have been done adequately for that language.) Translations that use TEV or FC as a base and stay fairly close to the form of those versions, can attain 60% work. When careful work is done using RSV as a base and TEV as a model to find good independent solutions for the target language, scores of 80% or higher are reached. But additional independent, creative work is needed for a translation team to reach into the 90% range. To a large extent, the key to attaining a translation of exceptionally high quality is the ability on the part of the translators to identify problems systematically and to apply the procedures that lead to good solutions. Translators that are able to identify translation problems are already at least half-way to solving them. It is for this reason that we have worked on the classification of translation problems, and encourage consultants to include problem oriented checking methods. REFERENCE MANUAL FOR TRANSLATION PROBLEMS We are working on preparing a manual that will list, define and classify all possible translation problems. Examples will be given, and possible solutions illustrated. Classification can be achieved in various ways; we are hopeful that the one proposed will be the helpful. (We acknowledge with gratitude the input from other colleagues, particularly some important insights from Noel Osborne.) We are also proposing a check list of items that need to be checked during the course of working with a project--items dealing not only with translation problems, but also with text and manuscript completeness (thus linking up our work with the manuscript department). See the "Translation Quality Check List" attached to this paper.

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES

PAGE 5

Translation problems have been classified into the following general categories (a more detailed listing is given in the appendix, "Classification of translation problems"): SOURCE LANGUAGE PROBLEMS TEXT PROBLEMS EXEGESIS PROBLEMS (Interpretation) SOURCE-RECEPTOR LANGUAGE TRANSFER PROBLEMS GRAMMATICAL PROBLEMS LEXICAL PROBLEMS FIGURATIVE LANGUAGE PROBLEMS STYLISTIC PROBLEMS DISCOURSE PROBLEMS ORTHOGRAPHIC PROBLEMS SOURCE CULTURE BACKGROUND PROBLEMS RECEPTOR LANGUAGE PROBLEMS THIRD LANGUAGE PROBLEMS RECEPTOR LANGUAGE SUPPLEMENTARY MATERIAL RECEPTOR LANGUAGE MS FORMAT PROBLEMS TRANSLATION PROBLEM SOLUTIONS

Since all translation problems arise in the course of attempting to transfer the content of a source language text to a receptor language, most of the problems are grouped under the heading SOURCE-RECEPTOR TRANSFER PROBLEMS. However, some problems seem to focus more heavily on the source language while others on the receptor language. These are grouped under the headings SOURCE LANGUAGE PROBLEMS, RECEPTOR LANGUAGE PROBLEMS, and SOURCE CULTURE BACKGROUND PROBLEMS. Since many teams are not able to translate directly from the original languages, secondary language bases are needed (English, French, etc) as source texts. This gives rise to special THIRD-LANGUAGE PROBLEMS. Finally, there are other types of problems associated with translation work grouped under the headings RECEPTOR LANGUAGE SUPPLEMENTARY MATERIALS and RECEPTOR LANGUAGE MANUSCRIPT FORMAT PROBLEMS. Source language problems Source language problems have to do with text and exegesis (variant readings, alternate renderings). These must be solved before transfer can begin. For most translators, solutions are found in commentaries and in versions of the Bible that are particularly helpful as textual and exegetical bases. Textual and exegetical base versions put the expertise of scholars, who have grappled with the problems at hand, at the disposal of the team.

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES Source-receptor language transfer problems

PAGE 6

These are translation problems with a double focus; both the source language and the receptor language are in focus. In some cases, however, either the source or receptor language will be perceived to be in more immediate focus. The non-equivalence between languages causes a tension when the form and content are to be transferred from one to the other. This lack of correspondence between languages requires that adjustments be made to the form of the receptor language; these adjustments take the form of expansions, deletions, and other types of alterations. In general, the principle of the priority of content over form is applied. Problems in this category fall into the following subcategories: Grammatical Lexical Figurative language Stylistic Discourse Orthographic Grammatical problems have to do with obligatory alterations in the surface grammatical structure. Expansions, deletions and alterations attempt to provide grammatical equivalents in the receptor language. Lexical and figurative language problems deal with the linguistic form and its referent in the outside world. The deep structure of the meaning of the forms is in focus as semantic equivalents are sought in the receptor language. Culture background problems deal with the cultural disparity between the source language and receptor language. Some problems belong to more than one category, so decisions regarding classification may seem somewhat arbitary at times. However a classification system of this nature is needed as a handle to get hold of the various problem types for translation and teaching purposes, as well as for systematic checking. It would be helpful to have a data bank of translation problems in some central location. Consultants would send in their observations and these would be made available to others, and writers of handbooks, for example, would have ready access to a wealth of material. Translation problems could be made available in either Biblical order or grouped in translation problem categories.

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES HOW TO CONSTRUCT TRANSLATION CHECKING GUIDES

PAGE 7

Until there are adequate materials available to help systematize checking, consultants will have to continue to rely on their own resources. This section attempts to outline various approaches that are helpful in this regard. (1) The existing Handbooks can all be used fairly easily as checking guides with a minimum amount of preparation. The chapters or passages to be checked are selected and then the problems dealt with in the Handbooks for those sections are noted. Some consultants mark up their Handbooks to identify the problem types. Checking sessions with a project can be very productive when guided in this way. Consultants will, of course, be sensitive to other possible problems not mentioned in the Handbooks. (2) The indexes at the end of Nida's "Theory and Practice of Translation", and Wonderly's "Bible Translations for Popular Use", and other Bible Society and Wycliffe texts provide excellent material for checking sessions. The problem type can be identified alongside the reference in the index to make them more useful in checking sessions. (3) Another method is to use a base-model approach to checking. With an RSV and a TEV in front of the consultant as he checks a translation, a multitude of questions will arise as differences are noted between the two versions. The RSV reflects the original language base, and the TEV is an example of a translation that has applied dynamic equivalance principles; these point to possible problems and solutions in the translation being checked. There are other base-model versions that can be used in English. In Spanish RVR-VP and in French Segond-FC are good pairs. (See the Supplementary Materials for a discussion of these concepts.) (4) Loewen's drill books (including "The Practice of Translating") are another rich source of topics, already classified. (5) Some consultants bmark up an RSV Bible r where crucial problems occur, and thus have a ready list for checking. Many entries accumulate over the years, providing a rich source of checking materials. Most consultants do not ordinarily keep track of translation problems they have dealt with; so this approach has the added merit that a record is kept of some of the significant examples arising out of their experience. Various approaches to checking a translation will be used during the years that a consultant is in touch with a project. The degree of detail with which a text is examined will vary according to the complexity of the passage being examined and the amount of time that a consultant has available to spend with the project.

UBS TRANSLATION WORKSHOP '84 15-26 MAY 1984, STUTTGART TITLE: TOWARD SYSTEMATIC TRANSLATION CHECKING PROCEDURES

PAGE 8

Following are some aditional comments and suggestions for variations in checking procedures: (1) The examination of a lot of connected text, verse by verse, is important for all projects. It is essential for checking discourse features. (2) For some types of checking (grammatical and lexical matters) a sampling procedure of looking at every 5th (or 10th) verse can help one cover a lot of ground in a short time. (3) Checking that concentrates on a list of specific problem types can be very helpful in pinpointing areas in which the translation team needs further training. The cumulative effect of checking the same problem in a number of passages is very effective in pointing out areas of weakness in translation knowledge. (4) It is also important to check the translation of words that have more than one meaning . The contextual translation of key Biblical words needs careful checking in many contexts. UBS translation consultants usually are not able to check the translation of every verse of a translation. That is left to exegetical experts on the team. However, by using careful random sample methods, we should be able to cover a lot of ground and gain deep insights into the quality of a translation. Systematic checking techniques are essential for the development of a good translation team and for a proper assessment of the quality of a translation.

Extracted text from https://www.dropbox.com:

Download Fehderau Trans Checklistfrom Robert Bascom (UBS Global Mission) NameSizeModified 01Fehderau p1-14 paper.docx36.59 KB42 hrs ago 02Fehderau Checklist appendix.docx24.96 KB42 hrs ago DATA 1 Examples.docx19.57 KB4 days ago DATA 2 Examples.docx19.52 KB4 days ago DATA 3 Examples.docx34.85 KB4 days ago Fehderau Trans Checking Original.doc169.22 MB4 days ago Flag for copyright Loading