Journal article Open Access
Haspelmath, Martin; Siegmund, Sven
In this short contribution, we illustrate what a typological replicatio would look like by examining some of Greenberg’s (1963) word order generalizations in a range of different 30-language samples. This is not a real replication based on new data (which would have required a substantial amount of research, beyond the scope of this discussion note), but a “simulated” replication in that it is based on data gathered and published independently: Matthew Dryer’s word order data published as individual chapters of the World Atlas of Language Structures (WALS). Our little exercise illustrates what replication in typology would look like, showing that it is not only possible in principle, but prospects for replication are actually quite good. Greenberg (1963) used neither sophisticated sampling techniques nor thorough in-depth analyes of all the languages concerned. And yet most of his results are replicable and have been replicated.