([Blodgett and Schneider, 2018](/bib/blodgett_and_schneider_2018/)) detail the application of this scheme to English possessive constructions: the so-called **s-genitive**, as in [exref 001 Genitives/Possessives], and **of-genitive**, as in [exref 003 Genitives/Possessives]:
- [ex 001 "<u>the Smith family[p en/'s Possessor]</u>/<u>[p en/their Possessor]</u> house ([ss Possessor])"]
- [ex 002 "<u>the tea[p en/'s Gestalt]</u>/<u>[p en/its Gestalt]</u> price ([ss Gestalt])"]
- [ex 003 "the house [p en/of Possessor] the Smith family ([ss Possessor])"]
- [ex 004 "the price [p en/of Gestalt] the tea ([ss Gestalt])"]
Note that the s-genitive is realized with case marking (clitic [p en/'s] or possessive pronoun[^1]) rather than a preposition, and the case-marked NP in the s-genitive alternates with the object of the preposition in the of-genitive.
(This may feel unintuitive: annotators looking at the s-genitive construction are often tempted to focus on the role occupied by the head noun rather than the case-marked noun.)
The s-genitive and of-genitive are particularly associated with [ss Possessor] (which applies to a canonical form of possession) and the more general category [ss Gestalt]; both supersenses are illustrated above [exref 001 Genitives/Possessives], [exref 003 Genitives/Possessives].
In addition, both genitive constructions can mark participant roles and other kinds of relations, including [ss Whole] and [ss SocialRel] relations.
When the s-genitive is used, the <i>function</i> is always either [ss Gestalt] (most cases) or [ss Possessor] (when the possession is sufficiently canonical).
While overlapping in scene roles with the s-genitive, [p en/of] is considered compatible with some additional functions, including [ss Whole], [ss Source], and [ss Theme]; thus of-genitives with such roles do not need to be construed as [ss Gestalt] or [ss Possessor]:
- [ss SocialRel--Gestalt]:
- [ex 005 "the grandfather [p en/of SocialRel--Gestalt] Lord Voldemort"]
- [ex 006 "<u>Lord Voldemort[p en/'s SocialRel--Gestalt]</u>/<u>[p en/his SocialRel--Gestalt]</u> grandfather"]
- [ex 007 "the hood [p en/of Whole] the car ([ss Whole])"]
- [ex 008 "the nose [p en/of Whole] He-Who-Must-Not-Be-Named ([ss Whole])"]
- [ex 009 "<u>the car[p en/'s Whole--Gestalt]</u>/<u>[p en/its Whole--Gestalt]</u> hood ([ss Whole--Gestalt])"]
- [ex 010 "<u>He-Who-Must-Not-Be-Named[p en/'s Whole--Gestalt]</u>/<u>[p en/his Whole--Gestalt]</u> nose ([ss Whole--Gestalt])"]
- [ex 011 "the arrival [p en/of Theme] the queen ([ss Theme])"]
- [ex 012 "<u>the queen[p en/'s Theme--Gestalt]</u>/<u>[p en/her Theme--Gestalt]</u> arrival ([ss Theme--Gestalt])"]
- [ex 013 "<u>Shakespeare[p en/'s Originator--Gestalt]</u>/<u>[p en/his Originator--Gestalt]</u> works ([ss Originator--Gestalt])"]
- [ex 014 "These are children[p en/'s Beneficiary--Possessor] clothes.[^2] [clothes intended for use and possession by children] ([ss Beneficiary--Possessor])"]
The literature on the genitive alternation examines the factors that condition the choice of construction; important factors include the length and animacy of the possessed NP.
In addition, [p en/of] participates in certain constructions that are not really possessives—e.g. <i>this sort [p en/of] sweater</i> ([ss Species]).
Some difficult cases are clarified below.
## Person in relation to a place or travel
In relation to an act of travel, the person is treated as a (possibly non-volitional) participant in a motion event. Otherwise, a person in relation to an associated place is [ss Gestalt].
- [ex 015 "[p en/my Theme--Gestalt] <u>destination</u>/<u>journey</u>/<u>travels</u> ([ss Theme--Gestalt])"]
- [ex 016 "[p en/my Gestalt] <u>hometown</u>/<u>birthplace</u> ([ss Gestalt])"]
## Idioms
Certain idioms require an s-genitive argument that does not participate in any transparent semantic relationship; for these, [ss `$] is used ([ss `$]).
[^1]: For ease of indexing, [p en/'s] or [p en/s'] is preferred over possessive pronouns for s-genitive examples in this document.
[^2]: Cannot readily be paraphrased with [p en/their] because <i>children</i> is not referential, but rather refers to a kind. This construction has been termed the <i>descriptive genitive</i> ([Quirk et al., 1985](/bib/quirk_et_al_1985/), pp. 322, 327–328).