Mining Dialogues for Variation



Following on from a previous post about understanding adjacency pairs and the usefulness of language-learning dialogues, I thought I'd share some strategies for using dialogues to arrive at the kind of semantic breadth you want to attain in order to handle common interactions fluently.

As discussed before, a single utterance may elicit dozens of possible responses. "Dozens" is rare, though — most of the time, there are only a few possible responses, particularly to common exchange types. For example, a greeting or farewell may have only one possible response or it may have 3 or 4 possible responses. Regardless of the total number, you will need not only to know all the possibilities, but be able to produce them automatically and without thought.

The first step is to understand the communicative process (and the notion of adjacency pairs discussed in the previous post). Then, you want to gather and learn to comprehend all the possible interactions within a given context. In other words, learn all the greetings and all their responses; learn all the ways of thanking a host, and all the ways of replying. These ritualistic functions comprise a tiny overall portion of a language, but simultaneously comprise a high proportion of frequency. More importantly, they build comfort and rapport on both sides of the conversation and are an essential component of communicative competence. A violation of an adjacency pair — intentional, unintentional, or simply brought about by an uncomfortable hesitation — will be very jarring to the native speaker.

For many common languages, it's easy to find language learning dialogues. A single dialogue will contain some useful information, but by matching up as many dialogues as you can find, you can begin to get a sense for:

  • Which expressions are used most frequently
  • Variations of set expressions
  • Possible variations in adjacency pairs

This can be very useful, and is one of the main reasons for using multiple texts and sources — unless, of course, you have a single resource that provides this collation of material already.

In essence, you can "learn a language" from any resource that provides text, translation, and audio recording. Just having those three things dumped in front of you, however, leaves a lot of the work up to you. If you know what to look for and how to use the materials, however, you can find gold where others see dirt. This is one of the key values of the essential linguistic and language learning training provided in our core course.

If you already have several resources for a language and haven't yet done so, take the time to match up interaction types and look for repeated patterns and variations. This process can be a goldmine for establishing the automatic responses necessary to gain instant credibility with native speakers.

Back to blog