Universal #1079

Legacy number: 1079

 Number  1075
 Original  If case, and especially genitive, is marked (enclitically) at the end of NPs, then word order is mainly SOV, NP Postposition, Genitive Noun, Noun Adjective (Greenberg’s type 24).
 Standardized  IF case, and especially genitive, is marked (enclitically) at the end of NPs, THEN word order is mainly SOV, NP Postposition, Genitive Noun, Noun Adjective (Greenberg’s type 24).
 Keywords  case, enclitic, order, adposition, attributive, adjective, noun
 Domain  inflection, syntax
 Type  implication
 Status  achronic
 Quality  absolute
 Basis  Cashinahua (Pano), Wappo (Yukian), Diegueño, Mohave (both Hokan), Basque, Tunica (both isolates), Kuuku-Ya÷u (Pama-Nyungan), Classical Tibetan (Sino-Tibetan), Oromo, Dasenech (both Cushitic, Afro-Asiatic), Hare (Athabaskan) all type 24 (SOV/Po/GN/NA), but also Japanese (Japanese-Ryukyuan) (type 23: SOV/Po/GN/AN), Tonkawa (Hokan)(type 16: SVO/Po/GN/NA), English (Indo-European) (type 10 or 11: SVO/Pr/NGvGN/AN)
 Source  Householder 1988

← Back to Universals Archive

Discussion
Counterexamples
Notes
What is an issue here in the implicans is the right-edge-marking subtype of phrase-marking of case, as opposed to head-bound word-marking of case (as in flexive-type morphology). It might seem surprising that English is not considered a counterexample. But then Householder appears prepared to weaken the implicatum to “THEN (a) there are at least two cases and (b) the language is predominantly suffixing and (c) there are at least some subordinate clauses with V last or some NPs with adjective or other qualifier last” pg. 385. Householder also suspects a link with ergative alignment, on the strength of this implication (p. 386, rephrased): IF there is ergative alignment, THEN grammatical relations are either marked on the verb (agreement, cross-reference), with NPs being uninflected, or by case suffixes.
Comment
Frans Plank · 3 Aug 2006
What is an issue here in the implicans is the right-edge-marking subtype of phrase-marking of case, as opposed to head-bound word-marking of case (as in flexive-type morphology).
It might seem surprising that English is not considered a counterexample. But then Householder appears prepared to weaken the implicatum to "THEN (a) there are at least two cases and (b) the language is predominantly suffixing and (c) there are at least some subordinate clauses with V last or some NPs with adjective or other qualifier last" pg. 385.
Householder also suspects a link with ergative alignment, on the strength of this implication (p. 386, rephrased):
IF there is ergative alignment, THEN grammatical relations are either marked on the verb (agreement, cross-reference), with NPs being uninflected, or by case suffixes.