Public Services and Procurement Canada
Symbol of the Government of Canada

Institutional Links

 

Important notice

This version of Favourite Articles has been archived and won't be updated before it is permanently deleted.

Please consult the revamped version of Favourite Articles for the most up-to-date content, and don't forget to update your bookmarks!

Search Canada.ca

A Procedure for Self-Revision

Brian Mossop
(Terminology Update, Volume 15, Number 3, 1982, page 6)

Revision is an essential stage in the translation process, but it does not need to be done by a "reviser." While new translators should have their work revised by someone more experienced, once they have shown themselves capable of achieving a certain standard they can begin to perform the revision function themselves. But it is then no longer sufficient for them simply to "go over" the draft, in the way they did while working under a reviser. Instead, they must devise for themselves a definite self-revision procedure.

It is perhaps true that revision by a second party—a "fresh eye" with no personal commitment to the draft—is somewhat more likely than self-revision to lead to the identification of problems in the draft. But revision by a second party is very time-consuming because the reviser must independently work out the overall argument and conceptual difficulties of the text. Much time will also be lost if the reviser does not distinguish necessary changes from changes that merely bring the draft into conformance with his or her personal style of writing, and the translator will waste time thinking about whether to translate in a certain way merely to please the reviser. These problems do not exist in self-revision, and the disadvantage of not having a second opinion can be overcome to some extent by the use of a method designed with this problem in mind.

The purpose of revision is to increase translation quality, but it must be distinguished from the quality-control procedure used by some employers, translation schools and professional associations for purposes such as hiring, marking and admittance to membership. First, whereas quality control is a procedure for evaluating a completed translation, revision is an integral part of the translation process. This is especially true in self-revision, in that the translator can divide translation tasks in various ways between the drafting and revising stages. Second, the quality-control procedure for identifying problems may be very elaborate and time-consuming because the work of several different individuals has to be evaluated on a fairly objective and consistent basis, or because a translation organization wants a detailed analysis of the weaknesses and strengths of its product. Revision does not have such aims, and there is a more limited amount of time available—time during which problems must be not only identified but also resolved. The approach must therefore be more subjective and less detailed.

The goal in revision is to determine the most important problems and resolve them. In other words, revision is not retranslation: if when revising you think that whole sentences of the draft need recomposing, then either you are seeking a degree of perfection impossible to achieve in the time available, or else there is a serious problem with the way the translation was originally drafted—something which cannot be remedied by revision.

When revising a draft, the question to ask is not "Can this be improved?" (for any piece of writing can be improved with enough time and effort), but rather "What needs to be improved?"*

Revision is basically a type of editing. Webster’s New World Dictionary defines "revise" as ’read over carefully and correct, improve or update where necessary’ and "edit" as ’revise and make ready for publication.’ The editing of draft translations does not always include the whole process of "making ready for publication," and in the case of translation as opposed to other types of writing, "updating" is irrelevant and "correcting" includes the correction of mistranslations. But otherwise the revision stage of translation has the same purpose as any other editing procedure: to look at the draft as a composition, written in a particular language, that must conform to certain writing standards and be "receivable" by prospective readers.

The need for revision of drafts arises from two aspects of the drafting process that make it difficult at that stage to see how the text will appear to prospective readers:

(i) While drafting the translation, the translator moves through the text relatively slowly and composes only a small fraction of the English translation before turning back to the French original. This fraction may get longer as the translator becomes more experienced, and the translator may work more quickly, but it will still be difficult for most of us to fully grasp the overall flow of the English text while it is being composed.

(ii)While drafting, the translator must inevitably read the French first (since no English as yet exists), with the result that the meaning is coming into the translator’s mind from the French text, whereas the prospective reader will have to get the meaning from the English text.

The revision procedure which I will now outline is designed to deal with problems arising from these two aspects of the drafting process. It does so by focussing on the English text: is the translation a coherent and genuinely English composition, and does it convey the desired meaning on its own, without there being any need to refer to the French? What the procedure specifically avoids is any repetition of the drafting procedure: first looking at a sentence of French, then turning to the corresponding English. It thus allows the translator to take a fresh look at the English, seeing it differently from the way it was seen during drafting.

As far as I know there is no generally recognized procedure for revising. What follows is simply an adaptation to self-revision of the procedure for revising others which I have worked out for my own use over the past several years. Other procedures suiting individual translators are of course possible, but I think they must all be designed to deal with the problems that I have just mentioned.

Step 1

Read a couple of paragraphs of the English without looking at the French.

Comparison of original and translation is a part of revision (see Step 2), but revision cannot be reduced to an exercise in text comparison. Only by reading the draft without looking at the French—reading it not as a translation but as English—is it possible to determine whether the draft is a coherent and truly English composition, and how it will impress its Anglophone readership.

During such a reading, certain problems of language and style can be identified: incorrect spelling, grammar and punctuation (which includes underlining and paragraphing); poor English usage; wrong level of language; gallicisms; language unsuited to the genre; poor intersentence transitions; excess verbiage; wrong focussing; wrong pronoun references, and uneuphonic effects.** In some cases, problems can be corrected as soon as they are found; in others, correction must await comparison with the French (in Step 2).

Three examples will illustrate the value of reading the English without first looking at the French:

(i) excess verbiage

If you have just read "les renseignements qui se trouvent à la page 17," you are less likely to notice that your draft translation "the information which is found on page 17" probably needs streamlining to "the information on page 17."

(ii) language unsuited to genre

If you have just read "acidifier chaque prélèvement (1 goutte CIH concentré)" you are less likely to notice, especially if you are new to scientific translation, that "acidify each sample (with 1 drop concentrated CIH)" needs changing to "add 1 drop concentrated HCI to each sample."

(iii) incorrect focussing, wrong paragraphing, and wrong pronoun reference

Suppose you have translated the following passage:

"Depuis une dizaine d’années, on parle beaucoup, surtout en Scandinavie, de l’acidification des lacs. Certains spécialistes attribuent ce phénomène aux pluies acides, tandis que d’autres proposent des explications différentes.

Pour éclaircir ce problème, à la demande des pays nordiques, une étude très importante a été entreprise dans le cadre de l’OCDE, par 14 pays de l’Europe de l’Ouest. L’évaluation des transferts a été réalisée au moyen de modèles numériques de transport basés sur la conservation de masse du SO2 à l’intérieur d’un certain volume d’air en tenant compte du temps de réaction du SO2 sur les autres polluants. Des modèles de type eulérien, puis lagrangien, ont été utilisés pendant plusieurs épisodes. [Ils nécessitent la connaissance de la hauteur de la couche de mélange; celle-ci a été évaluée par des mesures du SO2 et des sulfates par avion. Comme résultat de cet effort, on commence à avoir une idée assez bonne au sujet du transport du soufre fossile remis en circulation par l’activité humaine, cependant on a à peine progressé quant à la compréhension de la corrélation entre le soufre et l’acidité libre des pluies.

De vastes programmes de recherches (MAP3S et SURE) ont été récemment lancés aussi aux États-Unis d’Amérique, à ce sujet.

Le CO2 est une autre substance dont il est difficile de dire, s’il est un polluant ou non.] L’homme joue certainement un rôle dans la mobilisation du carbone fossile et dans sa présence accrue dans l’atmosphère. La circulation globale du CO2 est à peu près expliquée. Ce qui l’est moins, c’est son rôle dans le changement éventuel du climat…"

And suppose your draft of the bracketed portion of the passage reads:

"…To use these models, the height of the mixing layer must be known, and it was determined by measurements of SO2 Sulfur dioxide and sulphates taken from a plane. This study has begun to give us a good idea of the extent of the transport of fossil sulphur put into circulation by human activity, but there has been little advance in our understanding of the relation between sulphur and free acids in rainfall.

Large-scale research programs on this subject (MAP3S and SURE) have recently begun in the United States as well.

It is also difficult to tell whether CO2 is or is not a pollutant…"

If you are revising by going through the text making a sentence-by-sentence comparison with the French, and if you have just read the last French sentence in the bracketed passage, then the last sentence of the draft may appear to convey the right meaning. But in fact there is a problem. The sentence as it stands will normally be read with the following stress pattern:

"It is also difficult to tell whether CO2 is or is not a pollutant."

But this does not flow on coherently from what precedes. The argument is "as with SO2 Sulfur dioxide (just discussed), so with CO2." To get this meaning, the "also" of the draft would have to be read as going with "CO2" only, not with the entire expression "difficult to tell whether CO2 is or is not a pollutant." To obtain this result, the sentence would have to be read:

"It is also difficult to tell whether CO2 is or is not a pollutant."

But that is not the stress pattern a first-time reader will use. The problem becomes apparent if the sentence is read not in isolation after a reading of the corresponding French, but immediately after reading the preceding English. (Possible revision: "Another substance which may or may not be a pollutant is CO2.")

Besides revealing this focussing problem, independent reading of the above draft brings out the need for reparagraphing. A new paragraph should be opened at the sentence beginning "This study. . . ." As the draft stands, the study in question appears to be the study of the mixing-layer height, whereas in fact the reference is to the OECD study as a whole. (The paragraph should perhaps begin "The OECD study…"). The one-sentence paragraph beginning "Large-scale research programs…" should be made into the second sentence of the preceding (newly-created) paragraph.

Finally, continuous reading of the English reveals a problem of pronoun reference. The expression "on this subject" in the one-sentence paragraph could be interpreted as referring only to the "relation between sulphur and free acids in rainfall," a restriction which may not be intended. While the need for change here can be identified during Step 1, correction must await comparison with the French in Step 2, at which time the meaning of "à ce sujet" is checked. (Solution: delete "on this subject," to allow for both the restricted and more general interpretation of the topic of the "large-scale research projects.")

There is one type of problem which is not a matter of language or style but can nevertheless be both identified and corrected during Step 1. Since attention is focussed on the prospective readers rather than on the original author during this step, it may be apparent that certain things in the draft are unsuitable for the particular readership of the specific translation at hand. For instance, if you have just read, in a meteorology text written in France, that "la Météorologie nationale accomplit deux tâches principales," you will be under the influence of the French and therefore less likely to realize that your translation "the national weather service performs two main tasks" may not be suitable for Canadian readers: it may be necessary to write "the French national weather service."

Step 2

Read a sentence or so of the English, then look at the French and compare.

This is the test-comparison procedure, where omissions and mistranslations are identified in the chunk of the draft which has just gone through Step 1.

Judgments about the quality of the draft of a given sentence or expression can now be made in the light of the overall meaning of the paragraphs considered during Step 1 (the most common source of error in drafting is insufficient attention to context).

During comparison, the English is once again read first, so that the meaning comes from it, uninfluenced by any prior reading of the French.

Problems which are solvable fairly quickly are corrected as they are found. In more difficult cases, place a question mark in the margin and circle the problematic expression in pencil, for resolution in Step 3. Note that the paragraphs under consideration may resolve problems that were left question-marked in an earlier chunk of the text: an earlier problematic term may now appear in a new and clarifying context, or a problematic concept may be repeated in clearer terms. These terminological and conceptual problems may be ones that were not noticed during drafting, or they may be ones which were deliberately left for resolution during revision (more on this below).

Once sentence-by-sentence comparison is complete, Steps 1 and 2 are applied to the next chunk of text. Before proceeding, however, you may find it useful to reread the whole revised chunk. This does three things which you may not have dealt with during sentence-by-sentence comparison. (a) Verify that the revisions you have made fit into the flow of the text: no matter how good a revision looks by itself, it is not an improvement if it does not work in context. (b) Identify any language changes necessitated by a given revision: if in the sentence "The little enthusiasm and the lack of volunteers for an office baseball team are easily explained," the subject has been revised to "the lack of enthusiasm and volunteers," then "are" needs changing to "is." (c) Ensure a smooth transition to the next chunk, and terminological consistency within and between chunks.

After the whole text has gone through Steps 1 and 2, it may be of value (especially with long texts) to list, page-by-page on a separate sheet of paper, all the matters that have been question-marked in pencil during Step 2. This can be useful in ensuring that a problem on page 9 of the text and a similar one on page 39 are solved consistently. It can also facilitate discussion of problems when consulting the author or another resource person over the telephone. Finally, it can help organize the final resolution of problems in Step 3.

Step 3

For each unresolved problem, decide the priority and appropriate strategy for its resolution.

During both the drafting and revising stages of translation, it should be borne in mind that not all problems are of equal significance. It is especially important to be aware of this during the last step of revision, when final decisions must be made and time is of the essence. The translator should always be asking "What is important to my readers and what is not?"

Some problematic concepts and terms are more central to a text than others, and should be accorded higher priority for resolution (see strategies A and B below). Generally speaking, concepts are more important than terms (see strategy C). And depending in part on the genre and the purpose of the translation (eg. whether it is to be published or not), language and style problems will vary in importance: well-crafted sentences may be very important in the translation of a speech for public distribution but they are not so important in the translation of an inter-office report to be read by five or six individuals. In a scientific text, even one that is to be published, do not waste time pondering linguistic trivia ("as regards," "in regard to" or "with regard to") when conceptual problems remain.

There are six basic strategies for the final resolution of problems: (A) do more research (B) omit (C) invent an expression (D) fudge (E) place a question mark in the translation (F) correct the source-text author. Solutions C to F may call for a translator’s footnote, though footnotes should be avoided if possible.

(A) Do more research.

This is the most time-consuming strategy and is therefore applicable only to the central concepts and terms. The other strategies provide relatively speedy resolutions and are thus applicable to lower-priority problems, and in some cases to higher-priority problems where it appears that resolution by research would take an unacceptable length of time.

(B) Omit.

Where a concept is marginal or irrelevant to the subject of the text, and it is not clear how to express it in English, it can sometimes be omitted. For instance, I once translated a text on the causes of avalanches. The text began with a description of how the scientist arrived in a nice little Alpine village: the cattle were coming home and there were some pretty flowers growing in the meadows. The author then proceeded to name these flowers. At this point, I had left a blank space in my draft: as anyone who has translated biology texts will know, it can be extremely difficult to determine the right English common name of an animal or plant given only the French common name. Since this was not a biology text, however, I decided that the names of the flowers were irrelevant, and simply left them out. (Later, I deleted the whole passage about the village because its personal and anecdotal quality was foreign to comparable English scientific writing.)

(C) Invent.

If you have understood a concept in the French but cannot find the "official" English term, call a halt to your research and invent an English expression that conveys the idea. If you have found something which may be the English term, this could be added in a footnote.

(D) Fudge.

If it is not clear which of two possible meanings of an expression in the French text is correct, and there is an English wording which conveys both meanings, the problem can sometimes be left unresolved. Example: the deletion of the expression "on this subject" in the acid rain text discussed under Step 1. An alternative approach would be to express one possible meaning in the text and the other in a footnote.

(E) Question mark.

If you just do not understand an expression in the source text, do not pretend to have understood it. It is the mark of a professional to admit being stumped when she or he is stumped. Take a guess and enclose the expression in (?) question marks (?). A footnote may be added, giving a dictionary equivalent of the French (what some people call a literal translation) and perhaps an additional possible meaning. Question-marking is obviously undesirable in translations that are to be published.

(F) Correct source text.

An expression may be problematic because the author made a mistake. If you decide that this is indeed the case, simply correct the error in the translation (perhaps with a footnote explaining the original), or alternatively write [sic] in the text and add a footnote if necessary to explain the problem.

Having now described the three steps in my revision procedure, I should point out, before concluding, that what the translator does during revision depends to some extent on what he or she did during drafting. There can be various divisions of tasks between the drafting and revising stages of translation. For instance, in my own drafting practice, I do not try to solve all the conceptual and terminological problems that I find. Instead, for reasons that would take too long to explain here, I use the revision stage to solve some of these problems. Also, I like to move fairly quickly through the original composing of sentences. This means not necessarily including every aspect of the meaning of the French text, and not stopping to think of the mot juste if it does not come to me immediately. Instead, I concentrate on getting a thoroughly English-sounding sentence down on paper. Then I use the revision stage to bring the English closer to the French than my original draft, to the extent time permits.

Consider for example the following sentence from a Le Devoir editorial (Mar. 1/79): Il n’y a pas deux semaines, le même Globe and Mail pointait du doigt la presse francophone en lui reprochant d’avoir voilé les écarts de comportement du premier ministre québécois lors de la visite de M. Raymond Barre.

The following draft is acceptable:

Less than two weeks ago the Globe and Mail singled out the French-language press for not fully dealing with the Quebec premier’s bad behaviour during the visit by French prime minister Raymond Barre.

But (as a reading of the original context would confirm) the draft can be brought closer to the French by changing "singled out" to "was pointing a finger at," "not fully dealing with" to "glossing over" and "bad" to "improper." Note, in passing, that the need to translate "M. Raymond Barre" as "French prime minister Raymond Barre" (for Anglophone readers who may not know or may have forgotten who Barre is) was noticed during drafting in this case. If it had not been noticed, the resulting problem could have been caught during Step 1 of revision.

While many divisions of tasks between the two stages of translation are possible, I would counsel against any approach which leaves the elimination of Gallic syntax and style to the revision stage. My view is that once you have composed one of those awful sentences which are really French, but disguised in English words, it is very hard to get rid of them. You may not even notice them because, as their author, you have a certain personal commitment to them. Better to start with something thoroughly English, if slightly inaccurate, and then improve during revision.

I have mentioned that tasks can be divided in such a way that a certain amount of "writing" is left to the revision stage; that is, certain problems are deliberately left unresolved during drafting. It is also the case that a certain amount of "editing" can be done during the drafting stage. Changes can be made in a drafted sentence just after it is written down. Indeed, we all probably do a certain amount of mental "editing" before we write down a first draft. And theoretically, I suppose there is no limit to this. That is, it is conceivable that some people might have the mental ability to carry out the whole revision procedure either in their heads before they have written anything down, or else just after they have composed the draft of a sentence. In this case, no distinct revision stage would be required. The translator would sense what the finished text would read like while composing it. But this ability, if it exists, is probably very rare. It would therefore not be wise to set as a goal the elimination of the revision stage. However, the amount of work required during revision can be reduced, with practice and by experimentation with various drafting procedures.

Back to remark 1* Only in "training revision" would a reviser ask the first question, with a view to showing new translators how she or he would have translated a given passage.

Back to remark 2** In this article, my aim is to deal with the "how" of revision rather than the "what". For an interesting discussion (unfortunately without sufficient exemplification) of the types of problem one is trying to identify and resolve during revision, see Thaon and Horguelin, A Practical Guide to Bilingual Revision, Montreal, Linguatech, 1980.