Regulating Two at Once, II:
Congruence in ML+EL Constituents
and EL Islands
This chapter has several goals. First, it will continue the discussion from Chapter 4 of ML+EL constituents by introducing a new constraint on ML+EL constituents, a blocking filter. A Blocking Hypothesis proposes that a filter blocks not only EL system morphemes prohibited under the System Morpheme Principle presented in Chapter 4 but also EL content morphemes which do not show certain specified congruences with ML content morphemes. Second, in a related discussion, this chapter characterizes EL islands; obligatory EL islands are the result of certain non- congruences also involving the Blocking Hypothesis.
The System Morpheme Principle discussed in Chapter 4 makes the prediction that syntactically relevant EL system morphemes will not appear in ML+EL constituents; only syntactically relevant ML system morphemes are allowed. This prediction is supported by all available data; examples in addition to those already discussed will be presented later in this chapter. Clearly, this principle limits the occurrence of EL morphemes in these constituents. However, I now propose the Blocking Hypothesis as a corollary to the SM principle to limit further EL participation by blocking certain EL content morphemes as well.
In ML+EL constituents, a blocking filter blocks any EL content morpheme which is not congruent with the ML with respect to three levels of abstraction regarding subcategorization.
I use 'congruence' in this sense: two entities (linguistic categories in this case) are congruent if they correspond in respect of relevant qualities.