A-VISTA : NEW CHALLENGES FOR TAILOR-MADE TRANSLATION TYPES ON THE EXAMPLE OF RECORDED SIGHT TRANSLATION

With the growing market for language services, the changing needs of clients, and the technological advance, arise new tailor-made types of translation that fall in-between the traditional categories (such as oral/written, simultaneous/consecutive, transient/permanent, etc.), attempting to combine their advantages and best suit the given purpose. Initially inspired by a practitioner’s point of view, enhanced by a theoretical analysis, the paper presents the characteristics of a type of sight translation that is recorded for the client, which seems to be a very convenient form and an optimal solution for certain purposes.


Sight translation. Recorded sight translation. Definition of the terms.
Sight translation, also called on-sight and at-sight translation (in the US) or a-vista translation (in Poland, Italy, and other countries), is a type of interpreting/translating whereby the interpreter reads a document written in one language while translating it orally into another language.It is thus a hybrid form that falls between the categories of interpreting and translation, combining the features of both, probably closer to interpreting in the nature.
Sight translation in its standard form is nothing novel.It belongs to the three basic modes of interpreting, especially in the American classifications.It is used in aptitude testing for simultaneous interpretation; courses and exams in sight translation belong to the standard curriculum of training interpreters.Sight translation contests are organised in classical languages (e.g. by Eta Sigma Phi).It is also recognised as an official mode of legal interpreting.For example, The National Association of Judiciary Interpreters and Translators lists sight translation among simultaneous interpreting and consecutive interpreting -the three basic modes that have been adopted in federal and state statutes and court rules to be mainly used with forms and foreign language documents presented in courts.However, there it is recommended to afford the interpreter sufficient time to review the document's content before rendering it (Erickson et al 2006), which typically does not happen in the recorded sight translation discussed in the present paper.Also, United States District Court of New York adds that very densely worded documents such as contracts cannot be sight translated with any accuracy, but this technique is used to read relatively simple documents into the record (UDSC 2004), which is an important comment for legal texts.The present paper tries to show when sight translation of longer texts in a non-legal environment still is a better choice than written or other forms of translation, especially if the transient nature of the product is avoided.
The version of sight translation described in the present paper involves an additional element beyond the standard procedure.The oral translation is recorded into a Dictaphone -hence, such version is called recorded sight translation.
The recorded version of sight translation discussed here emerged from the specific demand on the market and belongs to the various non-standard language services.The client who brings a written text (usually a longer one) needs to understand its sense, the content.The text is mostly of an informative type (although one could imagine also an emotive or appellative text here), often a scientific article or report.An important suggestion from the client is the lowest possible cost and usually also short completion time.
The client often needs the information from the text to work on and use in his or her work.He or she is often a student writing a final paper, an academic worker or a professional undertaking some research, too busy to make long appointments with the translator, as might be necessary for standard sight translation, however, flexible in terms of the form of the translated text.The original language might not be known to him or her at all or -as in the case of English in Poland -might be known at an insufficient level to deal with a longer and more complex text.
In short, recorded sight translation is an optimal solution for low-cost, sense-oriented, fast translation of written texts that is not transient.

Translation procedure
The translation strategy of recorded sight translation may depend on the agreement with the client and the individual preferences of the translator.From my experience of a low-cost version, the most common procedure is that I do not read the whole text prior to interpreting it, only the abstract and summary, if present, to get a general idea.I review the text, and possibly check some key vocabulary in a dictionary or consult a specialist in the field.Then, I switch on the Dictaphone, and start reading aloud the translated version from the beginning of the text, stopping occasionally to allow myself read the coming portion of the text if it is necessary to get the meaning of the immediate couple of phrases (with long, complex sentences).Depending on the agreement with the client, I usually omit long references in the text if they are repeated (I usually just hint that there is one) and tables that are obvious to understand.I stress the titles, headlines, beginnings of sections, sometimes adding the page number.In terms of processing, I work at the level of a meaningful unit, a phrase, or a sentence, occasionally at the level of a section or paragraph.It is rarely necessary to stop the recording for a longer time, and look up a word in a dictionary or elsewhere.

Correction strategies
The correction strategies are various and their choice depends on the preferences of the translator and on the agreement with the client as to the "smoothness" (and price) of the final text.The choices include covert and overt correction.Covert correction, invisible for the client in the final product, means stopping the recording, rewinding it, and recording again the corrected word, phrase, sentence or greater portion of text.This option is more time-consuming, but does not disturb the flow of the text, and provides a more clear final version.It is a perfect choice for small, immediate corrections of vocabulary, syntax or general meaning, as well as for very serious mistakes that the translator feels are necessary to be corrected in this way.Overt correction, visible (or, rather, audible) in the final text, means correcting oneself without stopping the recording, as in "classic" interpreting.This option is less time-consuming, and is the best choice for obvious slips of the tongue or minor corrections that do not disturb the flow of the text.It might also be used as a rescue-strategy if the translator suddenly realizes a mistake made long before in the text and for technical reasons is not able to go back and re-record a large portion of the text.In such case, he or she may choose to record a remark about the correction.The latter strategy is typical for interpreting and is obviously not a perfect solution, but might be an optimal choice.

T h e o u t p u t
The outcome of sight translation is an oral draft text (like a working version), recorded on a Dictaphone in an analog or digital version, with the same meaningful content as the original.It is rather fluent, and may resemble the output of simultaneous interpreting or sight interpreting (simultaneous with an additional visual version of the original), e.g. in prosodic features.Thus, it is quite rough in comparison to a full written translation in terms of style, and is possibly also slower, with more pauses in comparison to a simultaneously interpreted version, especially with specialist or scientific texts with long sentences, as the translator must read what is ahead (plus, it apparently requires more effort and time to process a visual text than an oral one).Of course, it is also possible to stop the recording every sentence or meaningful unit, prepare a "smooth" version of the text, then switch it on to utter the translated text, and thus proceed by recording small bits of the text, resulting in a version almost identical with a written translation.The quality of such product is incomparably higher, but so is the cost.Moreover, a "smoother" version, although more time-and effort-consuming, might still be a bit faster (and thus, cheaper) than written translation, considering the average typing speed of a translator.

Theoretical characteristics of recorded sight translation
In a relatively basic theoretical approach to communication which involves translation, derived from the Jacobsonian framework, the three crucial participants are: the author of the source text (the primary sender), here mostly distant and not known personally, the translator (who becomes the mediator), and the client interested in understanding the message of the source text (the ultimate or terminal receiver of the message) (Grucza 1981, see also Kielar 1988).The communication process here may be seen as follows: the original message is expressed, here in the form of a written (paper or electronic) text, in one language.It is then read (or decoded), translated (processed), and read aloud (or encoded) by the translator, whereby the message channel is changed here from the visual to the acoustic (aural/oral).The client listens and, hopefully, understands the message.A possible model of the basic elements of sight translation is presented in Picture 1.
Adapted from the translation model by Grucza (1981) (see also Kielar 1988) As with all written texts, whether on paper or in an electronic version, the source text here is not transient, and thus may be used any number of times, by any number of people, at any time, and might be copied (technically).Moreover, it is not necessary for the communication participants, i.e., sender and receiver (author and reader) of the message to be present at the same time in the same place (spatial and temporal displacement) for the message to be conveyed.This form of communication might thus be very convenient, although more time and effort-consuming than the most typical oral one.
On the contrary, spoken texts are typically transient, momentary, and cannot be used again.They require the communication participants to be present at the same time and in the same place, or at least to be connected so that they may hear each other (e.g. by a phone line).This form of communication is generally faster and more efficient, provided it is easy for the participants to meet.
The transient nature of the spoken message may be a serious drawback.Various audio (and video) equipment solves this problem, and almost any oral message can be recorded in an analogical (magnetic tape, vinyl record, etc.) or electronic (digital) manner.Hence, the recording may be played more times, for various receivers in various places, might be copied, and it is no longer necessary for the sender and receiver to meet physically for communication to be possible.Needless to say, here we concentrate on the "verbal" content of the message, not on any other elements of Picture 1. Model of recorded sight translation communication and meaning otherwise conveyed in a personal meeting (e.g.included in the context, prosody, body language, etc.).
The interesting aspect of sight translation is that it uses a different channel of modality at the perception level (visual), and a different one at the production level (acoustic), and thus seems to require more mental effort (or at least, more cognitive mechanisms involved) than simultaneous interpreting or written translation.Nevertheless, this intuitive hypothesis has so far not been confirmed by empirical research.However, although the study of Lambert (2004) shows that sight translation actually yielded higher performance scores than simultaneous translation and even sight interpretation, the research focused rather on performance, not on the complexity of the cognitive processes.Moreover, it seems that the speakers' performance might have been enhanced by the additional preparation time, not the visual stimulus alone.(Lambert 2004)

Advantages
The advantages of recorded sight translation are many.From the client's point of view, it is usually the fastest way to obtain a translated version of a written text.Regardless of the type of text, sight translating typically takes less time for the same text than providing its full written translation.Hence, it is also the cheapest way to obtain a translation.Also, the client need not be present during translation, and even need not meet the translator at all or for more than making the order and receiving it.Moreover, the client is able to listen to the recording or its portions as many times as they wish, at any place, and the text may be duplicated.If the client already has some knowledge of the source language and of the subject, such translation often also helps them to advance their level of the language and enables them to deal with similar texts on their own in the future.
From the translator's point of view, it is very convenient to regulate the time, place, and speed of processing and producing the text (I personally prefer quiet cafés).Although it requires concentration and a lot of mental effort, sight translation is much less stressful than any other form of interpreting, as the translator has no immediate time pressure, is able to check certain vocabulary, make slight corrections and does not face the client while translating.It is also often acceptable if some vocabulary of the translated version is approximate as long as it conveys the message, allowing occasional double versions and minor additions as in simultaneous interpreting, which otherwise would be considered a mistake (see Hejwowski 2006).
Considering the translated product, it is of a better quality than with any other form of interpreting.Moreover, in terms of accuracy, sight translation is generally the same or very similar to written translation, (the accuracy level is very high), which has been proven by Sinaiko's research with professional conference interpreters translating the minutes of a U.N. meeting.He also observed that the sight translation was four times faster: 21 words per minute, as opposed to 5.4 (see Pierce 1966).Recent research by Moser-Mercer mentions even about 60 words per minute for beginners and 115 for professionals at sight translation (see Lambert 2006), although does not compare it with written translation, which has probably also risen.Even if we consider a recent improvement in both modes of translation, it still does not change the general dominance of sight translation here.
Another general advantage is more flexibility than typically with other types of translation.The quality of the final product is subject to individual agreement with the client.Depending on the preferences, it may be close to an "ideal" written text produced orally or rather a draft, working version, with various stylistic or prosodic imperfections, the latter version being naturally the cheapest, yet, fully understandable one.
Moreover, some clients may prefer an audio version of the text for cognitive reasons, if their dominating perception channel is acoustic, or for technical reasons, e.g. to listen to it during another activity, such as driving, walking or doing housework and thus to save time.

Drawbacks
Recorded sight translation is not suitable for all clients.An unquestionable weakness of recorded sight translation is the relative "imperfection" of the product, causing a greater demand for the recipient to process the resulting text: it is more difficult in perception than a written or freely spoken one.For example, the translated text may be difficult to follow for a person who is easily distracted by some possible clumsiness, an uneven flow of speech, pauses, approximations and other features typical for an interpreted text.However, every attempt to improve the quality of the product, usually by stopping or rewinding the recording, results in either greater cost on the part of the client or on the part of the translator (depending on the initial agreement).
Also, recorded sight translation is not suitable for all texts.Very specialised texts which require additional resources to work with, with a great proportion of specialist terminology, where it is best to use a computer and various search engines, or standard texts for which there are databases and software with matrix solutions (e.g.Trados) are obviously not good candidates, as it would be inefficient.
An obvious difficulty is the translator's relative effort (and thus, exhaustion), which is greater than in written translation (and only slightly or not at all lower than in any other form of interpreting) due to not only the synchronic mental processing of input and output, and the inconvenience of translating a written-style text into an oral-style one, but also to the change of input-output channel.
There are some additional general negative associations such type of sight translation might yield; for example, it seems like going one step back from quality translation, as the recorded text might be rather rough in terms of style or prosodic features.Thus, it makes an impression of a draft version (which it in a way is) or an unfinished product (which it is not).However, such option is sufficient for the client, with an optimal balance between the cost and the profit.

Recorded sight translation compared to simultaneous interpreting and written translation
It is difficult to classify sight translation (both recorded and unrecorded) into clear-cut categories.Certain features of the mental processes, the original, the product, the required skills of the translator and other aspects make sight translation more similar to simultaneous interpreting, while other ones bring it closer to written translation.Table 1 presents features of recorded sight translation that are in common with both categories.

In common with simultaneous interpreting
In common with written translation 1. the translating conditions: stress and effort some stress caused by: -high concentration during translation, great mental effort -trying to provide a translation as fast as possible (time pressure) -the fact that one is being recorded -not as stressful as typical simultaneous translation, as one is able to stop the recording, make corrections, and does not face the recipient information processing -simultaneous processing of input and output, mostly synchronic perception and production -processing smaller chunks of meaning and text -anticipation -the translation is produced relatively fast -focus on the "sense"; syntax precision not necessary -processing of visual information -occasionally also diachronic processing of input and output (e.g. at the level of a sentence), esp.when the recording is stopped and in cases of corrections corrections and hesitations -corrections "on the spot" (without rewinding the recording) -free comments may be added for emergency corrections -hesitation possible, as recording may be stopped -corrections by rewinding the tape and re-recording the text form of contact between client and translator -the client has contact with the translator's voice -the client need not be at the same time and place as the client during translation, -the client may never meet the translator resources while translating -relying mostly on the translator's immediate memory -it is possible to look up a word in a dictionary or elsewhere

the nature of the input (the original text)
-the original is a written text, typically in a printed (paper) version; the channel is visual 3. the nature of the output (the translated text) -the product is an oral text; the channel is acoustic -the outcome has some imperfections typical for an interpreted text (e.g. in style) -the translation is produced immediately, on the spot -intonation and sentence stress is an additional source of information -the product has fewer imperfections than a typical oral translation -the product is not transient, and thus can be used any number of times, in any place, at any time (it is lasting, portable, may be copied) by any number of clients 4. the skills required of the translator -good speaking abilities (clearly, evenly, with good intonation) -chunking information, anticipating, reformulating, handling difficulties on the spot -reflex, ability to concentrate, intuition, -good working ("operational") memory -self-discipline and selforganisation of time and place of work A characteristic feature of recorded sight translation, not typically present elsewhere is the flexibility of the agreement.It is up to the client to decide on the level of roughness, as it is related to the quality and cost.In general, if the client has no or little knowledge of the subject, and possibly also source language, the more "smooth" the product should be.And the reverse, if the client has good orientation in the subject and is able to follow the information in the text, more roughness is acceptable.
As a more general overview, it is possible to compare various types of interpreting and translation (excluding consecutive and media translations as irrelevant here) in terms of the presence of key features.Table 2 presents the selected types of relevant translation and interpreting, where the typical presence of a feature is marked with X.
As we can see from Table 2, the features of recoded sight translation make it a very good solution to balance the losses and profits of various types of translation, such as both synchronic and diachronic input processing, less stress for translator, more flexibility for both translator and client (displacement possible), with a higher quality of the product.

Table 2. Translation and interpreting types and their selected features. The typical presence of a feature is marked with X. The optional presence is marked with (X).
Inspirations, questions and challenges for the future Nowadays, with the development of modern technology, boundaries of many categories within them become fuzzy, recorded sight translation being one possible example.This brings about new challenges for the theory of translation, as well as for practitioners.
One of the challenges is to conduct research on the mental processes involved in sight translation.The subject has not been studied thoroughly, as it is difficult to find reliable tools for such research.
To assess the performance at sight translation seems easier.A well-designed study on the subject was carried out by Lambert (2006), as mentioned above.However, certain methodological drawbacks would require to repeat it.Her 14 subjects were given a 10-minute warm-up period to perform 5-minute tasks of sight translation and sight interpreting, but no preparation time was given before simultaneous interpreting, which makes the three cases not suitable for a comparison, as the variables were not equal.A valuable repetition might involve a study where all three tasks have the same preparation time and another one with no preparation time.Such results might be much more reliable.
The following questions on recorded sight translation do not yet have a clear answer: -Sight translation, although oral, is not called interpreting.Thus, where is the dividing line between translation and interpreting?Is it possibly in the input or in the output?(It seems that the input channel is critical in the present terminology.)A suggested shorter name for sight translation was also transterpreting (at a graduate seminar at the University of Warsaw); however, the proposed term -although "fairer", as it draws attention to the hybrid nature of sight translation -is not much shorter nor more clear than the established one, and might not be worth introducing.
-How to assess recorded sight translation: what makes a generally good one?How to assess the gravity of the mistakes?And if simply in terms of the amount of distortion they introduced into the text (see Mizon and Dieguez 1996), then how to assess that?-How to pay the sight translator: per working hour, per page of original or per minute/hour of the translated version?(The most objective seems the option per page of original.) Although sight translation is not a new invention, recorded sight translation is a simple innovation worth making popular.It is a perfect solution for a particular purpose, and is not practised as often as it would be profitable for both client and translator.Here the key role play translation agencies which offer various services and provide their descriptions -they could make it explicit that the option of recorded sight translation is available.
Sight translation also seems unappreciated, as it is associated with the traditional grammartranslation teaching method, a task type that everyone once had to perform in elementary school when learning a foreign language at a basic level, and is not considered a very complex cognitive operation.