Choosing RTT Test Points
In each survey, the choice of test points for administering Recorded Text Testing (RTT) will be influenced by the purpose of testing and the knowledge you have when planning the data collection trips. This knowledge can come from written sources, pre-trip interviews, local community interviews, and word list analysis.
People are different from place to place, even if they speak the same variety. So the places you choose will shape the data you collect and ultimately the conclusions you draw. When choosing test points, taking into account everything you know about the people and varieties will make the test more useful.
Sources to Consult when Choosing RTT Test Points
The sources below are the usual contributors to the first list of places where you plan to develop tests and administer them. The list may change as testing progresses, because you should always review it when you receive more information. The number of places you choose will depend on the time you have for the survey and the number of varieties you know about.
1. The purpose of intelligibility testing in the survey
Is it to provide a broad description of intercomprehension?
- This purpose will usually require the greatest number of reference tests, because each variety will need to be tested with each of the others.
Is it to identify the reference variety(ies)?
- This purpose will require fewer reference tests, because background research and interviews (both pre-trip and within the community) will be used to reveal the most likely candidates.
Is it to identify the peripheral lects that understand a reference lect?
- In this case, it is clear that one lect will serve as a reference, but it is unclear which of the communities that speak related lects understand it.
2. Pre-trip library research and background interviews
What have knowledgeable people outside the community said?
- Linguists, emigrant community members, NGO workers, and other people who have spent time in the community may have useful information on the names, locations, relationships, and comprehension of lects.
What do earlier surveys or censuses report?
- These sources of information may be useful especially for the names, locations, and perhaps population figures for the lects. You may be able to identify the boundaries beyond which testing is unlikely to produce helpful results.
What are the results of earlier intelligibility or comparative studies?
- These studies may reveal which communities ought to be included in testing solely on the basis of linguistic relatedness.
Where are the centers of commerce or government?
- These locations are highly influential and are easy to identify from maps and informal questioning at government offices and in the local community. They should always be considered for reference tests.
What geographic or political boundaries exist in the area?
- Such features can act as obstacles to contact and communication for closely related lects. Alternatively, by "fencing" communities in, they can act as catalysts for contact and communication for distantly related or unrelated lects. If boundaries seem to have an influence in the area, you should discover their effect on comprehension by including points within them in the tests.
3. Sociolinguistic community interviews conducted with speakers of the lects
In what places do they say the lects are spoken?
- The answers to this question, together with the information gathered before the trip, will provide a thorough master list from which to select places in which to develop reference tests and administer them.
Which lects or peoples do they associate or identify with?
- The answers to this question may help narrow the choices of reference tests.
How similar do they say the lects are to one another?
- The answers to this question will group communities and lects together and may help reduce the number of test points by lowering the possibility of doing more testing than is really necessary.
How well do the people say they understand the other lects?
- The answers to this question go beyond those of the one above and speak directly to the issue of comprehension. They may reveal that speakers of one lect understand another lect even though the two are not closely related, which is information that would not come out of a word list analysis.
4. Comparative analysis of word lists from all the lects
If word lists have been collected in each of the lects, then hopefully there are results from community interviews to accompany them that will expose the sociolinguistic information discussed above.
Which lects seem to group closely?
- Comprehension should always be tested among lects that share a large number of innovations. If a "parent" language still exists, it should be considered for a reference test.
Which sets of lects are not grouped together?
- Unless sociolinguistic information indicates otherwise, it is probably not useful to do comprehension testing between groups.
Which lects seem not to fall in any group?
- Use the results from community interviews to determine whether a lect is isolated or rather is associated with a lect in one of the groupings. If there is an association, then test speakers of the first lect with a reference test from the second lect.
