Cost in translation

Courts around the world are cracking down on costs yet the complexity of complying with disclosure rules is increasing, especially where multiple languages are involved. Drew Macaulay of Consilio looks at the major challenges facing litigators when dealing with multi-lingual disclosure

Macaulay: how do you know which languages are present?

Large companies today are more likely than ever before to operate on a global basis, dealing with parent and subsidiary entities, suppliers, customers and joint venture partners in other countries, aided by workforces that are able to communicate using more than just their native language.

This ability to do business across borders and in multiple languages brings significant competitive benefits. However, for companies faced with litigation or regulatory investigation, the existence of documents in multiple languages creates a major headache. Not only does this bring increased risk and delay to an already unpredictable investigation or disclosure process, but the additional costs can be significant at a time when corporate budgets are tight and courts in common law jurisdictions are expecting tighter cost control..

Deploying the right process and technology

The process of collating, searching, reviewing and disclosing documents is common to many forms of legal matter, from regulatory and internal investigations to litigation and arbitration. What is less understood is the additional complexity caused by the existence of documents in multiple languages and the risk they pose to legal practitioners.

It is relatively common to collate large sets of data (such as the e-mail archives of relevant company employees) and perform keyword searches across these data sets to identify documents which are potentially relevant to the matter at hand. If the data sets contain multiple languages, then any keywords that are language-specific need to be translated into all the relevant languages for an effective search to take place. This in turn raises another issue: how do you know which languages are present?

Continuing this theme, certain languages are written using non-Latin characters, for example the Hangul script is used for Korean and the Kanji characters for Japanese. If the methods used to collect and search documents in these languages involve computers which do not have the relevant language packs installed, a scrambling of characters occurs, again compromising the search process.

Speaking the same language

For a law firm undertaking a multilingual document review exercise the issues do not finish there. Many major law firms boast staff proficient in a range of languages, but they may not be available in sufficient numbers to complete the review process within the required deadline, and getting documents in a specific language in front of the right reviewer is not always straightforward.

Translation is the traditional method for dealing with this issue, but is frequently inefficient in this context as the process is very labour-intensive (and therefore expensive), and the precise relevance of a document has not yet been established. It is therefore quite possible to spend thousands translating documents which will be discarded as irrelevant almost immediately.

Thankfully, as the problems of undertaking multilingual document disclosure exercises have become more common, so have the potential solutions. In some cases just planning for the existence of these documents can mitigate significant risk. In others, more technical solutions may need to be considered. As with many problems, the best approach lies in breaking down the process into individual tasks, and using the most appropriate combination of human resources and technology to achieve the goal of each task.

Starting at the beginning, simply being aware of the risks that are posed by incorrect collection of documents will help enormously. In many cases, external assistance is used at the collection stage to guarantee the admissibility of evidence, and if there is a chance that non-Latin characters will be an issue in your case, it is worth enquiring as to whether your selected service provider has experience in dealing with these languages. Similarly, if the provider is going to be undertaking the searching of these documents, then confirmation that their software can handle searches in all the relevant languages should be sought.

If the languages present in the data to be searched are not known, then the use of automated language identification software can be extremely helpful. As its name suggests, this can automatically identify the language in which an electronic document is written, even where that document (such as a chain of e-mails) may contain two or more different languages. This enables the review team firstly understand which languages are present and in what numbers, and to then collect documents together by linguistic group for subsequent human review or translation.

Blending the efforts of automation and skilled staff

If reviewers with the required language skills are not available within the firm, then two further options to avoid the expensive and time-consuming human translation are the use of multilingual document reviewers or machine translation.

Machine translation is frequently a good option for documents in Latin-based languages, as the linguistic structures of these languages are often very similar, so the output can be reasonably accurate – at least enough to identify whether the document is relevant. With non-Latin based languages the quality may be less good, and it is usually worth asking the provider to undertake a test using a sample document before any final decisions are made to use this approach.

As an alternative, or where machine translation is not an option, it is possible to engage multilingual contract lawyers to undertake the review. The availability of contract lawyers proficient in rarer languages can however be a challenge, particularly at short notice, and it can be helpful to engage the services of a specialist sourcing firm with a large network of contract staff to avoid delays in getting the review up and running.

Determine the best outcome

Given the potential obstacles involved, the early implementation of the right combination of the above solutions for a specific case is essential. Once this has been accomplished, the application of modern workflow methodologies and technologies should mean that the process can be completed efficiently and accurately, ensuring that you and your clients remain on the right side of the courts.

Drew Macaulay is managing director at e-disclosure and managed review provider Consilio