Of Nominatives and Datives:
Universal Grammar from the Bottom Up*
Michael Silverstein
University of Chicago- Introduction: The lessons of phonological universale
That indefatigable pursuer of statistical generalizations, Joseph Greenberg,
asks (1978:1,6) why "phonology holds a special place in the attempt to
arrive at empirically valid generalizations across languages." He says one
gets the "impression both of precocious achievement and of coherent prog
ress when compared, for example, to syntax." He attributes the achieve
ment of phonology to "its relatively restricted and transparent subject mat
ter"; thus, "it serves to bring certain issues of explanatory theory into par
ticularly sharp focus." Perhaps so. But the lessons of phonology that ought
justifiably to be applied in the realm of "morpho-semantax," as we might
call the non-phonological planes of language, have not, to my knowledge,
been illustrated. Rather, over the years, one inappropriate formal concept
or method after another has been willy-nilly taken from phonology to other
linguistic planes, with dubious motivation and curious results. What I want
to clarify here, using some data on case systems that I introduced into the
literature a number of years ago, is the way in which we can transfer the
kind of precision of formulation of cross-linguistic regularities from phonol
ogy to syntax.
Where understanding of phonology has achieved some sort of univer
salist level of sophistication, namely, in the realm of segmental, word-level
patterning, there have been two major components to the underlying strate
gy: first, the Saussurean (or, more particularly, post-Saussurean) analysis of
segmental phonemes as differential, correlative, oppositive entities of sig-