Kayne (2005: 212) proposes the following relevant principle (see also Collins 2024 for discussion):
(1) Principle of Decompositionality (POD)
UG imposes a maximum of one interpretable syntactic feature per lexical item.
In this note, I will show how to derive this from more basic assumptions about Merge. Here are the assumptions:
(2) Definition: Merge(A,B) = {A,B}, where A and B are distinct syntactic objects.
(3) Definition: Syntactic Objects
a. Individual interpretable features are syntactic objects.
b. Lexical items are syntactic objects.
c. The output of Merge(A,B) = {A,B} is a syntactic object.
d. There are no other syntactic objects.
(4) Definition: Lexical Item
A lexical item is a phonological matrix associated with a set of syntactic features.
LI = {PHON, S}; PHON = phonological matrix; S = set of syntactic features.
(5) Axiom of Combination
Every combination (set) of syntactic objects is formed by Merge.
(6) Atomicity of Lexical Items
A lexical item is not formed by Merge of syntactic objects, nor does it contain any part formed by Merge of syntactic objects.
Now, given these assumptions, consider again the POD.
Suppose that we have a lexical item with two interpretable features F1 and F2. This would be represented as follows by the definition in (4):
(7) LI = {PHON, {F1, F2}}
But {F1, F2} is a combination of two features, each of which is a syntactic object. Therefore, by the Axiom of Combination (5), it must have been formed from Merge:
(8) Merge(F1, F2) = {F1, F2}
Therefore, by Atomicity (6), {F1, F2} cannot be part of a lexical item. Therefore, (7) is ruled out. Therefore, the POD is derived.
In the general case, any combination of two or more interpretable features would be a combination of syntactic objects. By the Axiom of Combination (5), it would be formed by Merge, violating Atomicity in (6).
One last comment: the form of a lexical item is a set LI = {PHON, S}, but according to the definitions, that is not formed by Merge because PHON is not a syntactic object. I leave it to future work to determine how PHON and S could be combined by Merge.
Discussion
The POD is very close to the Axiom of Combination plus Atomicity. The difference is in emphasis. The Axiom of Combination is much more general than the POD, applying to any combination of syntactic objects, so it has much broader empirical and theoretical scope. As for Atomicity, it is part of the definition of a lexical item, in the sense that syntactic objects are built by Merge from lexical items and individual features (generators), not the other way around.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.