What (else) depends on phonology?

Among the first implicational universals proposed were phonological ones such as, “If a language has nasalized vowels, it also has oral vowels”. However, as pointed out by Greenberg (1966), all (spoken) languages have oral vowels, so virtually anything can occur in the conditional clause if the consequent is an absolute universal (cf. “If a language has a velar implosive [ɠ], it has oral vowels” or even “If a language has a case system, it has oral vowels” (Greenberg 1966:509n)). I will start my discussion with some generalizations about phonological dependencies, in part to show that they can teach us certain things about problems in stating implicational universals. I will then turn to the long-standing attempt to correlate tendencies for phonological properties to correlate with morphology and syntax, e.g. “If a language has vowel harmony, it is agglutinative; if it has vowel reduction, it is flectional.” While most of these “hopeful” universals are either false or have not been shown to meet current standards in linguistic typology (Plank 1998), they indicate an interesting array of “motivations”, which will receive my attention: What caused Gabelentz, Bally, Lehmann, Dressler, Donegan & Stampe, Gil etc. to propose such correlations? Can phonology determine grammar? Can grammar determine phonology? Is there any tendency for aspects of each to covary? If so, why? One of my favorites, which I will discuss, is whether serial verb languages tend towards short words and tone (Foley & Olson 1985) plus I will offer a grammar-phonology dependency of my own! All of this will be presented within the context a continued investigation into the question of why (certain) post-Greenbergian typologists have so little interest in phonology (Hyman 2007).