Specializations > Morphosyntax

Books on syntax

<< < (2/3) > >>

Daniel:
As a rough summary, Formal Semantics piggybacks on Generative Syntax: the idea is to figure out how meaning combines within the syntax trees (so the "other" side of the trees, in a sense). ("Formal Syntax" could be a terminological substitute, but "Generative" is more common. Confusingly, "Generative Semantics" is something else, a distinct offshoot of an earlier version of Generative Syntax.) For the most part, Formal Semantics is concerned with intricacies of logic and compositionality, rather than specific details of the syntactic structure per se (although a lot of important and fundamental questions about syntactic theory have been addressed with respect to Formal Semantics). For example, semanticists might not make a distinction between "NP" and "DP" (a determiner phrase, which contains a noun phrase, e.g. "the red book" is a DP, containing "red book" an NP), because they're not so concerned with that level of detail in syntactic structure and instead concerned primarily with how the denotations of those elements combine. In short, you know know a bit about Generative Syntax (I'd recommend Carnie's book, or if it isn't too technical as an intro, Müller's), and the rest should work out from reading about Formal Semantics itself. (If you're looking for a textbook about formal semantics, I'd recommend Kearns' Semantics, which I've used to teach and is useful, but there are a number of books out there. Formal Semantics is a very technical topic but also inherently logical [that's the whole point], so you may or may not find it intuitive, but at least it is consistent. There is still some variation in approaches to semantics, roughly parallel to the variation in syntactic theories, but I'd say somewhat more navigable.)

Matt Longhorn:
I picked up Kearn's and it is really helpful.
A question on functional application where you have a one place predicate
if t is the sentence/proposition at the top of the tree and e is the argument node, the predicate is listed as (e,t). Is this a combination of the argument e with the t from the top of the node? or is it e combined with a newly introduced t?
In Dynamic Syntax I have seen this listed as (e -> t), presumably with the same meaning

Daniel:
(e -> t) seems like a clearer notation to me.* The comma-notation is not intuitive, but refers to the same input-output relationship. The predicate "SLEEP(x)" maps entities ("x") to whether or not they are sleeping (which is a truth value).

(*It is less convenient to type using special symbols, and also not as convenient for a shorthand when multiple levels of functions are embedded within each other. And it is also confusing because an arrow like -> is a technical symbol for 'material implication', a kind of logical conditional relationship between truth values.)

Matt Longhorn:
do you know what the annotation def is likely to mean? alpha = def alpha V up arrow, up arrow
Alpha is the greek letter not the word i typed, and up arrow are actual arrows

def = definition?

Daniel:
Definiteness and definition are possibilities. I'm not sure without context. I'm guessing alpha is some kind of variable (think "x" in algebra). In phonology it's used as a placeholder for features on a segment (e.g. voicing), while in syntax it may be used as an element in a tree (e.g. "Move alpha" or "Merge (alpha,beta)").

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version