Languages I / Universals and Typology
Practice question · Select all that apply

Why might a universal induced from a sample of languages be less secure than it looks? Select every legitimate concern.

Hints
  1. Three are the standard methodological controls; one is a blanket dismissal.
  2. Greenberg chose 30 languages deliberately for spread, and the results have largely held.
Show the answer
  • A. The sample over-represents well-described and written languages
  • C. Areal contact can spread a feature across unrelated neighbours
  • D. Genetically related languages may share a feature by inheritance rather than by constraint
Why

The first three are exactly why typological samples are stratified by family and region. The fourth overstates: Greenberg's 30 were chosen for genetic and areal spread rather than convenience, and the implicational results have largely survived samples fifty times larger.

Read the lesson: Universals and Typology →

Practise Universals and Typology

The app has 5 more questions on this lesson, and keeps your place in the course. Languages I is free to start.

More questions on Universals and Typology