Universal 1646: dependent-marked clause ⇒ dependent-marked relativization;
head-marked relativization ⇒ head-marked clause

Universal 1646: dependent-marked clause ⇒ dependent-marked relativization;
head-marked relativization ⇒ head-marked clause

Original

For dependent-marking patterns, clause marking type implies relativization type; while for the head-marking patterns, relativization type implies clause marking type.

Standardized

IF there is dependent-marking clause morphology, THEN relativization will be of the same pattern.
IF relativization is of head-marking type, THEN the clause uses the same pattern.

Keywords

head-marking, dependent-marking, relative clause

Domain

inflection, syntax

Type

implication

Status

achronic

Quality

statistical

Basis

Abkhaz, Adyghe (both NW Caucasian), Arabic (Semitic), Bantu family, Basque, Burushaski (both isolate), Chechen (N. Caucasian), Dyirbal (Pama-Nyungan), Hopi (Uto-Aztecan), Hurrian (isolate), Japanese (Japanese-Ryukyuan), Kaititj (Arandic, Australian), Lakhota (Siouan), Mayan family, Navajo (Athabaskan), NE Caucasian, Quechuan family, Tewa (Tanoan), Tonkawa (Coahuiltecan), Tungusic family, Turkic family, Uralic (eastern & western), Uto-Aztecan, Washo (Hokan), Yuman family

Source

Nichols 1984: 537

Counterexamples

Hopi (Uto-Aztecan), Kaititj (Arandic, Australian) have dependent-marking clauses but headless relatives, which are head-marking. Hopi is known to be under areal influence, The case of Kaititj is unexplained (Nichols 1984: 530).