If a language has case affixes on nouns, they are almost always suffixed.
Standardized
IF there are case affixes on nouns, THEN they are almost always suffixed.
Keywords
case, noun, affix-order, suffix
Domain
inflection
Type
no genuine implication; rather: provided that
Status
achronic
Quality
statistical
Basis
23-language sample in Sanders 1972; Hawkins & Gilligan surveyed 200 languages drawn from three samples: a 113-language sample of Stassen’s, a 40-language sample of Bybee & Perkins, and a 50-language sample of Gilligan’s; 50-language sample of Maxwell 1979
1. This is in essence what is being claimed: Case affixes on nouns tend to be suffixed. There is an implication hidden here, though: If an affix on nouns is a prefix, then it is almost never a case marker. 2. See also #1401 for Sanders’ claim.3. For typologically and diachronically motivated proposals for explaining why case suffixes are common but case prefixes rare, see Kahr 1976, Reh 1986. Generally on the suffixing preference see Hawkins (et al.), Bybee (et al.). An open question for these general accounts is why the suffixing preference is particularly strict for case, certainly stricter than for number. At any rate, there is a diachronic side to the universal, since what is at issue (in all accounts) is how free or clitic relational markers are getting morphologically bound.
1. This is in essence what is being claimed: Case affixes on nouns tend to be suffixed. There is an implication hidden here, though: If an affix on nouns is a prefix, then it is almost never a case marker. 2. See also #1401 for Sanders’ claim.3. For typologically and diachronically motivated proposals for explaining why case suffixes are common but case prefixes rare, see Kahr 1976, Reh 1986. Generally on the suffixing preference see Hawkins (et al.), Bybee (et al.). An open question for these general accounts is why the suffixing preference is particularly strict for case, certainly stricter than for number. At any rate, there is a diachronic side to the universal, since what is at issue (in all accounts) is how free or clitic relational markers are getting morphologically bound.