Sunteți pe pagina 1din 4

A Brief Look at Intersective Properties1

Nick Danis
Rutgers University
May 25, 2015
Following Alber and Prince (in prep.), Alber et al. (2015), and much recent work in modern,
rigorous OT, a typology is analyzed as a number of intensional properties. Each property "takes
on a set of mutually exclusive values, where each value is a ranking condition" (Alber et al.
2015: 4). Properties also have the possibility of being under the scope of another; that is, the
choice of PropA.x/y is only relevant when PropB.w/z is equal to z. The term relevant here refers
to whether or not the ranking condition defined by that property makes a distinction of optima. It
is also possible for a property to only be relevant when it is under the scope of two individual
properties. This is the subject for discussion here. I will call this type of property an intersective
property.
For an example of a scopal relation between properties, the values of the Ftyp.ia/tr property of
the system nGo (Alber and Prince in prep.) do not make a distinction between optima if that
language is specified as HavF.nil: if there are no feet, the ranking between foot type constraints
is irrelevant. Thus, the languages that specify Ftyp.ia/tr are a proper subset of the languages in
nGo; one property is under the scope of another. In the terms of Alber and Prince (in prep.), the
Ftyp property is moot with respect to the property HavF.nil.
Some systems have properties that are under the scope of two properties; these multiplydominated properties are intersective properties. One such example is the property of Complex
Voicing in system KP0 in Danis (2014). This property decides the fate of voiced complex stops
(e.g. [gb]), but the grammar must first both admit both complex stops in general and preserve
voicing distinctions. Other examples of intersective properties have been independently found in
versions of the syllable theory with a split Max constraint (MaxC and MaxV) and certain
metrical systems in Alber and Prince (in prep.) (Alan Prince p.c., class lectures).
A very simple (but perhaps not the simplest) system, called IP0, whose typological analysis
contains an intersective property is defined here. The definition of the system is defined solely
with hypothetical constraints, inputs, candidates and assumed violation marks. Thus by definition
below is also the universal support of the system.

Special thanks to Eric Bakovi for organizing this collection, and to Paula Houghton, Hope McManus, and Natalie
DelBusso for comments and discussion. And a very special thanks to Alan Prince, to whom this squib is dedicated,
and would not have been possible without. Author contact: nick.danis@rutgers.edu

(1)

VT for IP0
input output
/A/
a
b
/B/
c
d
/C/
e
f
g
h

M1 F1 M2 F2
0
1
0
0
1
0
0
0
0
0
0
1
0
0
1
0
0
1
0
1
1
0
0
1
0
1
1
0
1
0
2
0

The factorial typology (via OTWorkplace Prince et al. 2013) of this system contains 5 languages,
shown below.
(2)

Factorial Typology of IP0


Inputs /A/ /B/ /C/
a
c
e
N.N
a
d
g
N.Y
b
c
f
Y.N
b
d
g
Y.Y.N
b
d
h
Y.Y.Y

The typology has three properties, called PropA, PropB, and PropC, to mirror their relation to the
candidate sets. The language names given show the settings for each property, in that order. The
property definitions are given below, in the notation of Alber et al. (2015):
(3)

Properties of IP0
Name
Value
PropA.Y/N
F1 <> M1
PropB.Y/N
F2 <> M2
PropC.Y/N
F1 <> M2

If the above properties were able to combine freely, the resulting typology would contain 8
languages, yet IP0 contains only 5. This is because PropC only distinguishes optima when PropA
and PropB are both Y. This is clear when looking at the crucial ranking conditions for each
languages.
(4)
Rankings for non-Y.Y lgs
Lg. N.N
Lg. N.Y

Lg. Y.N

All three of these languages have at least one property (PropA or PropB) whose setting is N.
When this is the case, setting the values for PropA and PropB is all that is needed to define the
ERC sets of these grammars; the choice of optima for candidate set /C/ is predictable from this
ranking information.
When F1 M1 (PropA = Y) and F2 M2 (PropB = Y), the choice of optima for the /C/
candidate set is no longer completely predictable; it could be either [g] or [h]. The choice must
be resolved via ranking F1 and M2; this ranking is moot with respect to all other languages, as
none of the rankings in (4) crucially rank these two constraints. However, when PropA and
PropB are both set to Y, the choice must be made:
(5)

Rankings for Y.Y lgs


Lg. Y.Y.N

Lg. Y.Y.Y

Each of these languages crucially ranks F1 with respect to M2, thus each has a defined value for
PropC. These are the only languages with such a distinction, as PropC is moot with respect to all
languages that are not both PropA = Y and PropB = Y. Thus, PropC is relevant for those
languages which are the intersection of the PropA = Y languages and the PropB = Y languages.
The following represents the properties and their possible values hierarchically. Scopal relations
of properties are represented via domination, so the intersective property PropC has two
immediate dominators in the tree structure.

(6)

IP0 Property Tree


System
IP0

PropA

N
M1 >> F1

PropB

Y
F1 >> M1

Y
F2 >> M2

N
M2 >> F2

PropC

Y
F1 >> M2

N
M2 >> F1

In summary, an intersective property is an intensional property of a typological analysis whose


values only help distinguish optima when the values for (at least) two other independent
properties meet a certain condition. In the hypothetical case here, the values for PropC.Y/N only
distinguish optima when PropA = Y and PropB = Y. Such property relations have also been
found "in the wild", in typologies modeling segmental identity (Danis 2014), syllable structure
(Split Max syllable theory), and certain metrical typologies (Alan Prince p.c., class lectures).
References
Alber, Birgit, Natalie DelBusso, & Alan Prince. 2015. From Intensional Properties to Universal
Support. ROA-1235.
Alber, Birgit, & Alan Prince. in prep. Typologies. Ms. Universit degli Studi di Verona and
Rutgers University.
Danis, Nick. 2014. Deriving Interactions of Complex Stops. Rutgers University. ROA-1220.
Prince, Alan, Bruce Tesar, & Nazarr Merchant. 2013. OTWorkplace. Retrieved from
https://sites.google.com/site/otworkplace/

S-ar putea să vă placă și