The issue of TM ownership is progressively becoming a matter of controversy among translators, agencies and clients, and we feel it calls for serious consideration. The typical scenario would be as follows: the translator has created a memory containing the translation pairs as a result of a translation job, as commissioned by a client. This client will certainly get the translation and pay for it, but who owns the generated memory? The translator, since it is the outcome of his/her effort? The client, as the originator of the translation and its final user? What is more, if the same memory is used across different clients and different translations, would this exchange be subject to confidentiality codes?

We launched these questions and some more to catmt@egroups.com, a mailing list devoted to translation and technology, and it raised a wealth of interesting opinions. This article is the result of a joint effort to share views and opinions. The article is a digest of answers received from CATMT subscribers. Names and e-mail addresses of people who took part in the discussion are provided at the end of the article.

There was general agreement among participants in the discussion on the fact that the product, i.e. the translation, belongs to the client. Nevertheless, when it came down to the by-product, i.e. the translation memory data resulting from this translation work, different players claimed ownership: the producer of the source text, the actual translator -as the originator of the target text- and any intermediary in the process.

In this respect, Rochelle Blumenstein pointed out that the owner of the translation memory is the producer of the source text and advanced the following two reasons:

The opposite view was summarized by Danilo Nogueira as follows:

Both sides provide equally solid arguments for TM ownership, and depending on what party you sympathize with, your point of view will be closer to one or the other. In any case, it seems everything comes down, in the end, to the economic aspect, i.e. who is going to make profit out of this data. In this sense, and before we go on to the next section, we think it could be worth taking some time to consider how little thought this issue was given before CAT tools were introduced into the market. In a way, we could say that translation memory data have always existed some way or other, either in well-organized paper cards, on small pieces of paper scattered around the office, or why not in "brain format". Now it seems the problem is that the memory can be sold as a separate tangible product and everyone wants a share of it.

Pedro Satué Ripoll entered the discussion and stated that the "the translator should deliver just what he is being requested to deliver", which is what most of the participants thought. Apparently, this would solve the problem by making a distinction between TMs as productivity tools and TM data as a commodity. If the client wants not only the translation but the generated TM data as well they should state so in the translation contract and pay both for the translation and the aligned data.

This is what Steven Dyson thinks in this respect:

This takes us, in turn, to the issue of selling TM data to other translators or agencies, besides your direct client. Is this ethically right? Should, then, agencies distribute data among their translators or even sell it?

We enter here a sloppy path and, as Victor Dewsbery says, knowledge in our heads is ours and there is no way you can deliver it to others. You can deliver products using that knowledge (translated texts, word lists or style guides) and you can "use the knowledge gained by working for one client when you work for other clients, too. In fact, there is no way you can erase this knowledge and start from scratch". In this sense, we might say that the intellectual rights of TM data belong to the translator and that "in using this knowledge, you must use your sense of professional responsibility to avoid betraying any confidential information about client 1 when working for client 2". This is what we could call internal use of TM data, and everybody agrees on the fact that it is ethically right.

On the other hand, there is what we could call external use, i.e. the reuse of the data by a third party that neither produced the data (the translator) nor paid for it (the client). The exchange of data for money in this context arises all kinds of ethical questions, since it is difficult to determine which part of this data belongs to you and which to your client. Gudmun Areskoug summarizes this issue as follows:

Suzanne Topping reminds us that:

As we see it, the legal dimension will be critical in the following years. If TM data is a commodity and can generate revenue, either in the form of money or of increased productivity and quality, sooner or later there will be a market. In this context, clear legal regulations will help this new business; while the absence of guidelines, either legal or de facto best-practices, will seriously hinder new developments.

In our opinion, this new market will grow and be beneficial for the world of translation. This idea is supported by different intuitions and facts.

In this context, we strongly see the need for legal advice.

We decided to submit a message to the LegalTranslators list legaltranslators@egroups.com. In reply to a message similar to the one posted in CATMT, we got only one reply from Alexandra Scott, "a trade mark agent, not a lawyer".

The scarcity of answers, just one, is probably the most significant fact to be outlined. However, this is not exceptional. There have been similar attempts to dig into copyright and translation memories in a list called CNI-COPYRIGHT devoted to the discussion of any issue connected to copyright. A similar question to ours had been previously posted by Barry Caplan and his submitted didn't receive replies.

His message can be found at:

In this message, there is a reference to an article by Sue Ellen Wright (and an excerpt is included) about terminology databases. This article was originally published in the ATA Chronicle. A reference and a short abstract, not the article itself, are available at:

With TMX as a standard for data exchange and the growing demand for a faster translation, TM ownership will still remain a controversial issue for some time. In the absence of general rules, we are compiling relevant information in different section of TransRef.org:

We plan to update this section regularly with new opinions and findings. Your contributions are welcome, specially if they reflect your own experience or common practices.

Compiled by
Antonio S. Valderrábanos
TransRef Editor
editor@...
TransRef homepage

With contributions from CATMT and LegalTranslators subscribers:

CATMT and LegalTranslators are hosted by egroups.com
catmt@egroups.com
CATMT homepage
LegalTranslators@egroups.com
LegalTranslators homepage