insights
Here you will find a list of everything I write on this blog that makes me think, “wow! i don’t know if i’ve ever heard anyone say this before!”
It is extremely likely that many of these things have been said by others before, and are even widely accepted, and I just don’t know about it. In fact, I’m sure that some of them have been said before. I just want to know which ones, and where!!
Ok, here’s the list so far:
The Sisterhood Principle: Uninterpretable features can only be interpreted by the interpretable features of their sister.
Pengli’s Law: Anything property of syntax that is not from Merge or phonology is from semantics.
The Feature Projection Rule: Functions project their features; arguments don’t. (this is to be slightly revised)
Defining Head, Complement, and Specifier: The head of a phrase is its function, the complement of a phrase is its first argument, and a specifier of a phrase is one of its other arguments.
The Runner’s Lung Hypothesis: Having a lot of control over your vocal tract and breathing mechanism is a prerequisite for evolving a learned communication system—almost all species with learned communication systems either fly or swim. Humans evolved good vocal/breathing control because we were stamina runners, and this satisfied the prerequisite.
The Overactive-Trunk Hypothesis: (This is not mine! Credit goes to Logan R Kearsley for making this connection!) Elephants also seem to have learned communication. Maybe elephants evolved good vocal/breathing control because they also used their trunks for a bunch of other stuff, like lifting heavy things and storing water.
The Baby-Talk Hypothesis: The human phonology/song system was primed for the introduction of chunked meaning because we were likely singing to our kids while telling them stuff.
Phonology and birdsong merge nodes of the same type (Symmetric Merge).
Labels correspond 1:1 with reverse tree depth, assuming node-hoisting (a node can be copied one level up without merging with anything—or maybe by merging with ‘NULL’).
Syntax and semantics merge nodes of different types (Asymmetric Merge).
Labels are determined at terminal nodes by primitive semantic type; parent nodes always take the label of their functional daughter.
This is because compositional semantics introduces an inherent asymmetry: functions vs arguments. Phonology and birdsong do not have this asymmetry.
The Symmetry Constraint: Asymmetry is a prerequisite for recursivity.
The Memory Bottleneck: (just a testable theory idek if i believe in!) Birds/cetaceans/etc. can’t merge asymmetrically because they don’t have enough computational power to hold one ph(r)ase chunk in their memory as a single object while adding another chunk to it. Humans developed asymmetric merge when we gained this degree of computational power.