Suppose we accept the following convention of "layering of features": (19) "When a rule of grammar assigns features to [a morphosyntactic representation], and that [representation] already carries specifications for those features, then (unless, of course, the rule is explicitly stated so as to "change" the features involved, rather than simply to add them), the result is not that the new features and the old merge within the same complex, but rather that a new layer of structure is created, taking the old feature complex as its base." [Anderson, S. R., 1977b, “On the Formal Description of Inflection”, in Beach, W. A., Fox, S. E., and Philosoph S., eds., “Papers from the Thirteenth Regional Meeting of the Chicago Linguistic Society”, University of Chicago, Chicago, 21f.].
- Anderson (2004a), a pag.256-257