RSS feed

Managing risks and resources: a down-to-earth view of revision

Tim Martin, Directorate-General for Translation (European Commission)1


Revision is not especially well understood, either as a concept or as an activity. The new European standard on translation services has helped to clarify matters but understandably fights shy of linking revision to functional, purpose-related considerations. Such fit-for-purpose principles are, however, a sufficiently broad and reliable peg on which to hang a revision policy. For organisations and individuals alike, the overriding requirement when providing translation services is to balance risks against resources. Deploying resources in downstream revision may not be as cost-efficient as upstream quality-assurance measures, and is only truly effective when accompanied by feedback.


Risk management, resource management, quality management, fit-for-purpose translation, EN Standard 15038

Revision is a deceptively simple notion. In translation circles we think we know what it is and, safe in that supposed knowledge, we tend to see it willy-nilly as A Good Thing. I shall suggest in this paper that the truth of the matter may be rather more complex. This will amount to arguing not just — perhaps obviously — that no revision is better than poor or unnecessary revision, but that revision is a valuable and costly resource best applied selectively.

What is revision?

First of all, what is revision? The following acid test may be helpful in answering that question. Does the statement “I have revised this translation and found nothing at all to change” sound odd or perfectly normal to you? Do you perhaps even consider it to be a contradiction in terms? Your answer places you ¾ forgive the caricature ¾  in either the Webster’s or the Petit Robert camp: neither of them bad places to be, but each offering a very different focus.

Webster’s defines revision as “re-examination or careful reading over for correction or improvement”, in other words an activity that may or may not lead to another, but one that has better quality as its objective:

revision = reading ?corrective action? quality.

The Petit Robert, on the other hand, defines it as “improvement of a text through corrections”2, in other words the end objective and the activity are virtually coterminous:
revision = quality corrective action.

This is a rather classic illustration of the difference between bottom-up and top-down thinking: either we decide what we are or aren’t going to do by looking at what we have in front of us, or we start with an a priori concept (improvement) and then do some work (revision) to get there. No doubt we’ve all come across revisers with the a priori (top-down) mindset. And no doubt we’ve noted that while they’re often brilliant, they’re also very expensive (and sometimes wasteful) in terms of resources, since they’re singing to no one’s tune but their own.

So, to answer our starting question, revision is checking to see if changes are needed and it is making the changes themselves: a doubleton concept, with something of the blank cheque about it3. What’s needed, to keep it within sensible and affordable limits, is a guiding principle, regulating not so much how it is done — since idiosyncrasies will always prevail — but what it should be setting out to achieve. As it happens, the new EN Standard 15038 on Translation Services (EN-15038 2006) goes some way towards doing just that.

Revision and EN 15038

A major asset of the new standard — which has not been without its critics among freelance translators — is its emphasis on clear definitions. Revision, it is careful to say, must be carried out by “… a person other than the translator …” (EN 15038, paragraph 5.4.3.), thus distinguishing it from self-revision (which it sensibly calls ‘checking’)4. Revisers themselves are required by the standard to “examine a translation for its suitability for purpose” (EN 15038, paragraph 5.4.3) and to compare the source and target texts and recommend corrective measures. The term ‘review’ is defined in paragraph 5.4.4 of the standard in the same way, but in relation to target texts only. It remains to be seen whether the term in that sense achieves broad currency, but the intention to clarify terminology is laudable.

Here, then, is our candidate guiding principle, “suitability for purpose”, spanning the conceptual gap between seeing what is needed (revision as a good look) and doing what’s deemed best (revision as a task performed) — and neatly reconciling Webster’s and Petit Robert in the process. I shall argue later that this principle is both necessary and sufficient as a framework for risk and resources management. But having agreed on what a reviser should set out to do — check against purpose — we must now answer a further fundamental question.

Why revise?

There are not really two ways about this: the main reason for revising a translation, whatever its level of sensitivity, is to eliminate any errors it may contain. Improvement in other ways is an added bonus, justified to a greater or lesser extent by the type of text and readership involved. The ‘four-eyes principle’ is a good one, as any translator will attest, but it can’t, in all economic honesty, be used indiscriminately.

But why such parsimony with revision resources and why such reluctance to raise the quality ceiling across the board? The reason is simple: it is usually errors and not infelicities that sink translations and their makers, and errors do not impact equally across all text types and purposes.

That being so, and because to err is human, one might expect the translation landscape to be at least moderately strewn with the debris of undiscovered errors and the fallout from related litigation. But, as Byrne (2007) points out in a closely argued article, that doesn’t seem to be the case. Citing also a US researcher who could not locate a single court case or ruling in which a translator was found liable as a result of a poor translation, and noting that his own research yielded similar results, Byrne nonetheless goes on to conclude that “…it would be unrealistic to interpret this lack of cases as proof that translators do not make mistakes or that the issue of translator liability is not something with which we should concern ourselves. […] ..the implications of substandard translations must be treated seriously.” (Byrne 2007: 2).

That seems to me a very sane and sanguine conclusion to draw. Perhaps the lack of litigation reflects translator skill and reviser acumen, but perhaps too it is just luck and lack of customer expertise. Either way, we should retain a healthy fear of translation errors in texts with a serious legal, political or commercial dimension — which account for most of what’s translated where I work and must surely account for a large chunk of most corporate and freelance workloads.

All translators and translating organisations will have their own ‘hall-of-fame’ collections of potentially serious errors, detected or not before the translation’s release. My own recurrent favourite, a wholly counter-intuitive translation from French, is the stipulation that official documents (e.g. tenders, forms, certificates) are to be submitted within “…a delay of x days or weeks”5. This, it must be said, is something that postal services the world over can achieve effortlessly without further encouragement from hack (or machine) translators. Reviser, where were you?

Fit-for-purpose translation: just good enough — or just good?

We have so far suggested what revision is, what revisers should be doing and why they should be doing it ― on the occasions that they are. Let us now try to place revision within a governing framework.

In Europe, the notion of ‘fit for purpose’ has certainly been tarnished by its recent exposure as a UK government slogan, but it is probably still the best and safest peg on which to hang a revision policy. And whatever one thinks about the working detail of the new EN standard, its formulators have almost certainly done the translation profession a service by casting the standard’s core concepts so firmly in the functional mould — a welcome departure after the ISO 9000 series.

So what exactly is fit-for-purpose translation, and why should it command such credence that a revision policy can plausibly be built around it? After all, it has the inescapable ring of a business motto, redolent of cheap-and-cheerful furniture kits. Is it more? I’d argue that it is.

Evidence that fit-for-purpose translation may be more than a soundbite or a management mantra is at hand from Google, whose hits on the phrase lead first to ‘terms of business’, either in charters or in actual translation contracts, and then to research articles and conferences. Crucially, the phrase occurs in both the ITI’s and the ATA’s model terms of business, which encourage translators to state formally that they will supply a product that is … ‘fit for purpose’.

And fit-for-purpose translation is not without its theoretical credentials (whether or not it needs them), contrasting with the nativist stream that stresses universals and fitting squarely into the functionalist, pragmatics-based stream that feeds into communication theory and corpus linguistics. Is that important? Only to the extent that it lends a little substance to what otherwise might seem a rather flimsy or ad hoc approach.

In light of the above, I would argue that fit-for-purpose translation, when applied systemically to a varied workflow, is a viable way of using translation and revision resources intelligently. It is not a second-class alternative: fit-for-purpose products are either very good or good enough, but never less. Above all, though, the fit-for-purpose principle is an invaluable yardstick against which to balance risks and resources.

Balancing risks and resources

Given the theoretical yardstick of 'fit for purpose', how in practical terms can one balance risks and resources as a corporate or individual translation provider?

  •  Corporate providers

The main concern of corporate translation providers, as risk-carriers, should be to match job to translator in terms of linguistic and thematic expertise and of ability to meet the deadline concerned without slippage or loss of quality, especially where single jobs are divided among several translators6.

  •  Individual providers

Once they have secured a client base, individual providers should confront their risk by making sure they don’t overstate their production capacity or expertise in order to obtain work. It is here, when one is working pretty much alone, that the risks-resources balance is at its most precarious and needs to be kept under constant review.

  •  Corporate + individual providers

Assuming that job assignment and acceptance have been carried out professionally and candidly, the risks-resources balance still needs close attention when it comes to deciding whether or not a second pair of eyes should be brought to bear. For corporates, revision may well become more and more a matter of course as and when the new EN standard gains acceptance, though they will doubtless continue to weigh the need against the known purpose of the text and the known reliability of the translator. For individual providers, the cost factor will loom ominously large, but it’s not unreasonable to surmise that in order to meet the revision requirements of EN 15038, informal pairings and collectives may begin to flourish more widely7.

We can perhaps sum up the above observations by recalling two of the principles that emerge from the Chesterman/Wagner dialogue aimed at bridging the gap between the ivory tower and the wordface (Chesterman and Wagner, 2002). The first ― “Never translate blind” (Ibid: 44) ― refers to the importance of ascertaining a translation text’s purpose via briefing and specifications, while the other ― “Never translate alone” (Ibid: 86) ― refers to the many and obvious advantages of working in what the dialogue calls “mutual cooperation circles”.

Nor, it must be said, should revisers themselves work either blind or alone. Revision should always be carried out in tandem or at least in close contact with the translator, while translators should make a point of signposting successfully solved problems in texts they know will be revised, so that revisers do not duplicate their work unnecessarily8. Risks indeed cut both ways: towards the client, because a poor translation can have serious consequences, and towards the supplier ― because resources squandered impact substantially on cost-efficiency.

Revision as part of quality management

We have seen that revision is a high-value, resource-intensive operation best deployed in a spirit of active risk-management. The new EN standard is right to assign revision a key role in guaranteeing the quality of translation services9, and very understandably chooses not to dilute its own force by entering into conditionalities or considerations of acceptable risk. Realistically, however, translation practitioners are likely to do just that, whatever their view of the standard.

But revision is only one means of ensuring translation quality, and perhaps not even the most effective, because brought to bear so late in the process. If one accepts that all processes — not just production processes in the strict sense — contribute to the quality of a final product, then quality control in the classical ex post sense of revision loses its pride of place and becomes just one of a number of possible measures. Some of these ― recruitment (or job assignment), training, the IT dimension (e.g. use of CAT tools) ― have traditionally been regarded as ancillary to translation proper, and of course are so in the most obvious sense. But in a less obvious sense they contribute very considerably to the quality-improvement endeavour so often associated with revision alone.

What binds these surrounding measures with a common thread is the notion of feedback. For revision alone is an imperfect art and can never ensure that an intrinsically bad product will be rendered flawless. Nor indeed should it be seen merely as a form of corrective action. Its real strength and investment value is as a feedback tool that allows its results to be channelled back into the whole cycle of translation production in order to eliminate or reduce problems at source. Only when that happens can one claim that risks and resources are well managed.

  • Byrne, Jody (2007). "Caveat Translator: Understanding the Legal Consequences of Errors in Professional Translation." JoSTrans 7, 2-24.
  • Chesterman, Andrew and Wagner, Emma (2002). Can Theory Help Translators? A Dialogue between the Ivory Tower and the Wordface. Manchester: St Jerome Publishing.
  • EN-15038 (2006). European quality standard for translation services. []
  • Martin, Timothy (2002). "Is Revision Always Helpful." Terminologie et Traduction, Issue 2, European Commission.
  • Mossop, Brian (2001). Revising and Editing for Translators. Manchester: St Jerome Publishing.

 Tim MartinTim Martin is a senior member of staff in the European Commission's Directorate-General for Translation. After seven years as a translator and reviser, eight as a unit head and a further four in charge of English-language support services and quality management, his current position is that of Internal Communication Officer. He has a postgraduate qualification in Applied Linguistics and a particular interest in seeing traditional translator profiles broadened to include such activities as authoring, editing and summarising.

Note 1:
The views expressed in this article are those of the author alone and do not necessarily reflect those of the European Commission.
Return to this point in the text

Note 2:
Translation of the original French “Amélioration d’un texte par des corrections” (Le Nouveau Petit Robert, éditions Le Robert, 2007).
Return to this point in the text

Note 3:
Uncertainty as to which of the two faces should prevail is perhaps behind the confusing array of would-be (but not-quite) synonyms increasingly used to express the revision concept, viz. cross-reading, checking, re-reading, proofing, reviewing, QC-ing, etc.
Return to this point in the text

Note 4:
Mossop (2001) rightly stresses how important it is for translators to integrate a check of the draft translation into their production process, and calls this step “self-revision”. His use of this term is understandable, since it deliberately suggests a systemic element very different from a simple read-through. But let’s be clear: ‘self-revision’, an essential part of translation, is not really ‘revision’ at all — and the EN standard is right to spell this out.
Return to this point in the text

Note 5:
Mistranslation of the French ‘délai’, which means either a period of time or a particular deadline-date.
Return to this point in the text

Note 6:
This practice has always been prevalent among corporate providers but is becoming ever more frequent as the pressure for rapid turnaround of high-volume texts increases. The new European standard’s emphasis on project management is no doubt a response to this development.
Return to this point in the text

Note 7:
As indeed may commercial groupings specialising in quality-control services.
Return to this point in the text

Note 8:
A list of similar practical tips on revision-management can be found in Martin (2002).
Return to this point in the text

Note 9:
Paragraph 5.4.3 of EN 15038 states baldly that "The TSP [translation service provider] shall ensure that the translation is revised".
Return to this point in the text