Translation agencies or Language Service Providers (LSPs) often talk about the quality of their Translation Memories, or TMs. So what is a TM and why is it important?

Put simply, just about every professional translator on the planet uses some form of software to speed up their work. Contrary to what you might expect, these software packages do not actually ‘translate’ the data fed into them (also this can be arranged by means of plug-ins, and is sometimes even what clients want). Instead, they allow the linguist to see both the original text and his/her translated work side-by-side and to make use of a TM.
Ah yes..what’s a TM?

Once a linguist translates a passage of text (SDL Trados, the most popular package calls them ‘segments’), the Translation Memory stores this for re-use as soon it encounters the same passage, or a similar one, at a later date.

This has massive benefits – the most obvious being that very repetitive texts (lists of data, for example) can be translated in a fraction of the time that would otherwise be required.

It also means, however, that LSPs can ensure clients a considerable degree of consistency – even over many years – in terms of vocabulary choice and style. Also, larger organisations such as vehicle manufacturers, will often have their own TMs which they will insist on all linguists using.

Combined with a “termbase” of pre-agreed terminology, a good TM allows linguists to offer unprecedented levels of speed while always guaranteeing consistency and accuracy.