Based on my experiences in the Stanford Universals Project in 1969-76, it seems to me that the substance and goals of language-typological research were the same then as they are today: the study of the distribution of structural characteristics across languages and in particular, the search for crosslinguistic similarities that are not due to genetic relatedness and areal effects.
The differences that I see are of two kinds: some aspects of methodology, and the place of language-typological research within the overall field of linguistics.
Although Alan Bell – a member of the Project – wrote a paper on sampling (published in 1978 in the first of the four volumes titled Universals of Human Language), most of the projects at the Stanford Universals Project were based on convenience samples. This contrasts with today’s considerable attention to sampling techniques.
2/ DATA SOURCES
We used grammars that were available in Greenberg’s and Ferguson’s personal libraries and in the Stanford University Library. We also used language consultants although sparsely. Corpora were not mined for data in any of the projects. The wonderful range of data availability that the internet now provides was of course nowhere there.
I am not sure exactly how we interpreted the concept of universals in the 70-s. However, it is clear that there was less awareness of the essentially unprovable universality of crosslinguistic generalizations. Statistical computations, commonly used nowadays, were not generally carried out.
B/ THE STATUS OF LANGUAGE-TYPOLOGICAL RESEARCH WITHIN LINGUISTICS
4/ RELATION TO GENERATIVE GRAMMAR
As I recall, none of the members of the Stanford Universals Project were generative grammarians. At the time, there was no wide-ranging crosslinguistic work done by Chomskian linguists and, since Chomsky had unkind words to say about Greenberg’s work, the Universals Project’s activities were viewed as sharply distinct from and even orthogonal to the efforts of generative linguists. This is clearly different now: generativists are doing thorough crosslinguistic comparison and forums such as Linguistic Typology or the ALT conferences do include generativist work. While in the 1970-s language typology was probably considered as an optional and fairly marginal field of linguistics, by now it has become an organic – and perhaps the most central – area of research in linguistics. This is also shown by the fact that, whereas in the 1970-s and 1980-s few academic programs offered courses in language typology, this is now common practice supported by the considerable number of textbooks on language typology.
5/ DISSEMINATION OF RESEARCH RESULTS
Due in part to the lack of the kind of technology we have today, the distribution of results was restricted and slow. We had a working papers series at Stanford and this is where most of our papers were published but it was fairly narrowly distributed. This has hugely changed by now, with the establishment of ALT and its journal Linguistic Typology and with many other linguistics journals also publishing crosslinguistic studies. Large-scale international projects as LINGTYP and WALS and the resulting publications have greatly contributed to establishing a central position to language typological research.