Japanese logogram for editing/proofreading

Checking and editing an existing Japanese or Italian translation to improve it.

translation (Japanese and Italian for us) is the only or among the only professions in which the resulting work product is routinely submitted to "another person" for verification under ISO standards. This can be justified by the need to ensure maximum accuracy in the translation of particularly sensitive texts. Sometimes, however, it simply becomes an unjustified additional cost for the end customer. More recently, under the name of MTPE, it manifests itself in a new role in close intimacy with the now ubiquitous machine translation.

“Editing and Proofreading” — Contents

Innanzi tutto, chi sono questi “terzi”?

In genere sono altri traduttori professionisti, ma a volte revisori specializzati che amano fare solo questo. Spesso sono semplicemente traduttori mediocri più o meno onesti lasciati liberi d'intervenire con correzioni che il traduttore originale non ha modo di vedere e tanto meno contestare. Il fatto che questo controllo sia così diffuso (e persino contemplato in una specifica norma ISO per fornitori di servizi linguistici (LSP) non significa però che sia sempre necessario.

Infatti, non di rado è superfluo e non fa che aggiungere un ulteriore onere per il cliente finale senza dare in cambio un vantaggio reale. In ogni caso, i concetti stessi di revisione testi e correzione bozze sono spesso fraintesi anche dagli addetti ai lavori, il che porta a una certa confusione nella loro attuazione. In Aliseo Japan™ abbiamo il nostro modo, crediamo corretto, d'intenderli.

Introduction to our Japanese and Italian editing and proofreading service

According to ISO standards and now a well-established practice at so-called LSPs, a translation should go through four distinct stages:

This procedure is now a widespread custom, but is doesn't mean that every translation must necessarily go through stages 2, 3, and 4. It all depends on the particular problems that could arise from not applying one or more of these stages.

As the primary link in the transactional chain, translators should use any means available to thoroughly check their works before delivery for contents, syntactic and grammar errors, and possibly strictly follow any style manual imposed by the client.

Such a check is not always sufficient, though. In practice, the more critical eye of a second person (e.g. a linguist dedicated to editing or another freelance translator) can detect the smallest of errors that manage to slip past the repeated checks of the translator. In other words, translators acting as reviewers possess a special knack for discovering the mistakes of others, although they may still fail to notice their own.

Needless to say, the scope of proofreading and editing should be extended to technical or specialized translation, legal translation, website translation and, why not, even to subtitle translation too.

Our editing and proofreading languages

  • Italian
  • Japanese

A generalist translator who is only qualified from a linguistic standpoint may face difficulties with certain specialized texts and therefore end up translating with an excess of creativity and guesswork; in such cases, specialized translators of the source language may be preferable due to their expertise in the subject matter, even if they lack extensive vocabularies or perfect writing skills. Texts produced by such translators should then be reviewed by a proofreader to render them more readable.

In most cases, a few translation errors do not adversely affect the content's usability or lead to accidents or legal troubles. In such cases, the actual usefulness of the review may be questioned, the most obvious effect of which is to unnecessarily raise the price for the end customer.

But sometimes these are real risks, making it prudent to have the translation checked by another professional to avoid situations where a serious mistake may result in injury or death—or simply damage the reputation of the client. Such checks are necessary because translators are only human and mistakes will happen occasionally, although it is important to note that professional editors are not infallible, either.

We can serve as translators or reviewers for Japanese-Italian translations. When translating, we deal with phase 1 (verification); when asked to review a translation, we handle either phase 2 (editing) or 4 (proofreading) or both, depending on the instructions received.

With all that said, it is important to fully understand the differences between editing and proofreading. Despite the fact that many professionals and translation companies use the terms interchangeably, they are not the same thing.

What is Editing?

what is editing?
Editing service

Also referred to as a bilingual check (or “cross check” or even native check in Japan), this process aims to correct and improve an existing translation by comparing it to the source text. Clients typically request editing to ensure that a translation correctly reflects the source content, especially in cases where inaccuracies, improper register, or a general unsuitability for the target audience could cause serious health and safety problems or otherwise reflect poorly on the client.

Consequently, editors must be very well versed in both the source and target languages and have access to the source text.

Oscar Wilde

Oscar Wilde

Writer, aphorist and poet

I was working on the proof of one of my poems all the morning, and took out a comma. In the afternoon I put it back again.

In general, editing should be entrusted to a professional translator or to an editor who is a native speaker of the target language. With highly technical/specialized translations that require the use of a particular terminology that can not be easily learned, however, at times it may be better to hire a native speaker of the source language who has a firm grasp of the specific concepts that appear in the translation. A proofreader can then be hired to check the text prior to publication.

Generally speaking, it can be said that reviewers should also be an indirect teaching tool for the original translators, so they should be fundamentally better at translating (writing style aside). But this happens less than it should.

Editing requires:

  • An editor/revise who is very familiar with the subject matter
  • An editor/reviser who is a native speaker of the target language but has also mastered the source language
Typically, checking a translation basically means vetting each single word and meaning, which can take even longer than translating from scratch because the editor/reviser is forced to work within a template that has already been defined by the translator, and this can end up being more of a hindrance than a useful reference. Therefore, editing should be paid based on the translation rate (this are ours just for reference) or on the actual time spent revising.

What is Proofreading?

Proofreading is a post-editing phase in which the translation is checked for correctness as a monolingual text. It does not involve the verification of terminology.

Proofreading service
What is proofreading

In addition, many translations require a solid multidisciplinary knowledge (often a mixture of mechanics, electronics, physics and/or chemistry, legal jargon, and so on), the acquisition of which for pure linguistically trained translators normally requires many more years of translation practice as well as being more difficult. We have a solid academic background acquired in various technical and non-technical fields in previous work experience.

Then, although linguistic proficiency is one of the fundamental requirements of any translator, what makes the real difference is the translation experience which, as such, requires time and constant practice.

Together with the perfect knowledge of the passive language, translation experience also helps to avoid too literal translation, the number one enemy of good writing style especially in translation from Japanese, which due to the substantial structural difference from English and Italian requires a special effort of stylistic reworking.

This means that the proofreader's mastery of the subject matter and source language is irrelevant. In general, the time required to proofread a text is directly proportional to the number of words that it contains, hence proofreading services are usually based on word count.

But clients often request proofreading for texts which really need a full editing pass first, a misunderstanding (unacceptable from translation companies) perpetuated by the lexical confusion between editing and proofreading. By the way, we also have a page dedicated to translation companies.

The reality is that translations produced by non-native speakers of the target language—a very frequent phenomenon—almost always contain far too many grammatical and syntax errors in addition to translation errors and a writing style that necessitates a complete overhaul, or even a new translation.

Proofreading requires:

  • A proofreader who is a native speaker of the target language

The use and abuse of editing

Translation is a highly creative profession and translators possess varying levels of expressive creativity.

Distinctive writing styles may be used to create variations of a translation that are all fundamentally correct. Ask ten translators to translate the same lengthy and complex text, for example, and you are likely to get ten different yet accurate results. If those ten translations are then sent to ten different editors or reviser, ten new versions of the translation will be produced, meaning that you will have your choice of twenty unique variations on the same source text! But which one is the "best" translation?

The risk of wrong editing
Employing a qualified editor is essential

The above example demonstrates the need for clients to clearly define the editorial scope. Doing so will prevent the editor's ego or desire to self-promotion (deliberately damaging the original translator) from compromising a translation which might already be perfectly usable.

Editors or reviser should ideally refrain from correcting a translation directly, instead leaving the task to the original translators and respect their right of objection. This unfortunately does not always happen; a client will usually adopt the new version under the assumption that its revision by another professional (not to mention the additional money spent) guarantees its accuracy, or because merely proposing corrections instead of actually implementing them will confuse the original translator, take too much time or be not practical.

In any case, end clients (but also translation companies) are seldom capable of or willing to evaluate a translation quality themselves—therefore neither the editing/revision quality—particularly when they don't understand the languages involved (which is quite often the case with Japanese). Haphazard editing can easily ruin a good translation and causing a waste of money and the original translators discarded because of the failure to evaluate them properly, which is precisely why they should always be made aware of any corrections made to their work.

Joys and sorrows of MTPE

MTPE means Machine Translation Post Editing. In essence, it means correcting machine errors, which in the field of translation are made by machine translation software.

Probably more than half of the world's population has heard of Google Translator in one way or another, while for “insiders” the preferred machine seems to be Deepl Translator nowadays, which deposed Google from the first place on the podium. There are other players, but they have to be content with a smaller slice of this very profitable market (except for translators).

It cannot be denied that at least between similarly structured languages (Italian and English, for example) machine translation has made substantial progress in terms of form and content, provided that the source text is relatively simple, structured with relatively short periods, in relatively well-known domains, with a controlled language, simplified syntax, reduced vocabulary and of course when the source text is perfectly native.

Still, caution should be always exercised because surprises are never lacking. Google Translator, for example, translates quite literally and in language pairs such as Italian-Japanese it's still not very useful (on the contrary it can be rather dangerous), while Deepl translates in a more natural way but often "tries to be smart", that is, omitting in the translation the parts that it doesn't understand, or adding on its own initiative words that are totally irrelevant. Try to believe!

Machine translation:
imposed by clients
or freely chosen by translators as additional linguistic resource?

There are two ways of looking at machine translation post-editing: that of translation companies, which use machine translation to provide translators with a "pre-translated" text that purportedly “only” requires some corrections here and there—therefore a job to be paid less, even much less than a traditional translation— and that of translators themselves, who may tend to consider machine translation more a technical aid to replace or add to the traditional dictionary and the omniscient Internet. The difference is substantial: to suffer the imposition by clients—whose aim is to reduce costs—or choose it freely as an additional tool to support the traditional translation work.

To be a little cynical we might say that the basic concept behind MTPE (or PEMT), or simply PE (Post Editing) as it’s often called too) as seen by clients is that the machine translates at no cost and human editors only correct errors and adjust syntax, thus earning significantly less than they what they would by translating from scratch, entirely with their brain.

But this editing takes longer, because unlike real translation, which requires to read only the source text—and from there to start tracing the basic semantic architecture—MTPE also requires to read and understand what the machine has translated automatically, the construction of which can, however, be very different from what the human translator instinctively prefers. After a first reading, the translator realizes that the machine has made real translation errors (which at a first might not emerge clearly due to its apparent translating quality) and has constructed the sentences in its own way, with fanciful periods and most probably by translating too literally.

After the first moment of uneasiness, the translator starts reviewing the terms proposed by the superior intelligence and does lots of copy-pastings to make the sentence more human, especially fluent and that does not seem to have been translated by a relatively good non-native speaker. Maybe the translator gets a jolt of nerves and deletes the entire machine translation to start from scratch because by dint of reading it and rereading it he cannot make heads or tails of it.

Maybe the translator even starts doubting that the poor machine could not do better because after all it was fed with a text written by a non-native speaker. Such cases abound more and more. Typically, however, the client won't be able to help the translator since he has no clue about the target language, maybe has other things to do and then, what the hell, he's paying after all, isn't he?

More often than not, clients request revision jobs suggesting that the translation was done by a real-life translator, whereas it was only machine-translated. By the time the text arrives, however, it may be too late to refuse the job, perhaps because translators don’t want to risk jeopardizing the relationship with the client whom—from time to time, fortunately—sends more “human” jobs. Or, when they find the strength to remonstrate, clients bring out the insidious idea of “light revision”: “Hey, we don't need all that quality. Just make the text comprehensible enough.”, except to be later bombarded by questions and comments from the final client who is not satisfied with the translation quality.

Some translation agencies try to convince the still hesitant translators by reassuring them of the excellent "pre-revision" work that their in-house specialists have done (“They have adapted the source text to make it more comprehensible to the machine, which translation as then been integrated with our 'reliable' translation memories. You will see that the work will be easier... and faster!”) The fact is, this is not always the case practically, especially when clients don’t know the target language at all. Sadly, this also happens with difficult or complex documents such as patents, court decisions, historical texts and everything that generally requires precision, research and/or stylistic reworking.

In essence, it can be said that for translators MTPE combines the disadvantages of machine translation with the difficulties of human translation, that is, the difficulties add up. End clients pay for the work almost if not as much as a human translation while getting something that can only be inferior to what good translators would do traditionally. This is because in one way or another they are led to follow the semantic structure aseptically traced by the machine and to accept too promptly and uncritically what it proposes unless, of course, they decide to do it all over again, but at that point working as much as they would have done without the machine’s help. That is, it is a compromise solution that as such satisfies only the translation company if involved, certainly not the translator.

Question for fellow translators:

Better being traditional translators
or
simple editors of
more or less intelligent machines?

Perhaps, one day machine translation will almost completely replace human translators, except for editorial translators, no doubt (or at least not in this century). While not yet reliable, many translation agencies and even more translators swear that they already can't do without it. We, on the contrary, prefer to continue to be translators, not correctors of more or less intelligent machines. We are happy to leave this new role to new generations of translators—provided that it will be still possible to continue to call them so—or to the more accommodating ones whose expectations are more in line with the secondary role that awaits them in the increasingly flat world of languages.

Editing and proofreading rates

Service Rate
Editing USD 40 €/h
Proofreading From USD €/page

Service terms

  • These rates are updated periodically to reflect changes in the JPY/USD exchange rate.
  • "Page" means 200 Italian or English words or 400 Japanese characters
  • The minimum rate per job for direct clients is USD 40.
  • Unless different terms are agreed upon in advance, all payments are due within 30 days of the invoice date. (Private individuals must make their payments before translation work begins)
  • Payment method: WISE, Payoneer, PayPal, credit card via the free PayPal service (account registration not required) or IBAN/SEPA bank transfer.

In the footer you will find a link to our Terms of Supply and Privacy Policy, which we invite you to read before submitting an editing or proofreading service order.


Notes on editing rates

Remarks on editing rates

Inexperienced or unqualified translators often appear as an economical choice to clients because they provide their services at such low rates. But when the resulting translation is fundamentally flawed or unsuitable for its target audience, lacks lexical breadth, has been done too literally, makes insufficient use of synonyms, or lacks fluidity, the client should accept the fact that the cost of its review can equal or even exceed that of the translation itself.

Reworking an existing translation often takes more time than re-translating from scratch since the editor is required to work within a framework that has already been predefined in some way, to maneuver carefully so that the correct portions remain intact, to adopt the original translator's mindset, and to some extent emulate the existing writing style. In other words, this process is much more mentally taxing on the editor and, paradoxically, increases the risk of introducing new errors.

There is no way to determine how long an editing job will take without examining the translation first. A rather common yet inexplicable (for us) notion is that the price of a review should be a fixed percentage (typically around 50%) of that spent on the translation. This doesn't work in practice because there is no real way to calculate a priori how many hours of work will be required without seeing the actual text.

In fact, the risk would be not only on the editors side because by failing to quantify their economic remuneration they would not earn enough with respect to the time actually spent, but to the entire value chain too, because by feeling forced to work fast and without the necessary commitment precisely because of the low editing rate accepted they would not carry out honestly the task received: improving the translation instead of worsening it.

Top