LECTURE 2.5 – Sentence Processing
Visual and auditory recognition is referred to as early processing whereas sentence
comprehension can be called late processing.
The debate between modular theorists and connectionists is, when do we start to use top-down
information? To answer this question we will need to consider many things.
Firstly is should be said that there are 4 stages to sentence comprehension.
o Lexical Access Recognizing words
o Syntactic Analysis Parsing a sentence
o Semantics What does this mean?
o Pragmatics What am I really trying to say?
The modular view argues for Autonomous Theory. They argue that these stages of
comprehension are autonomous in the sense that they do not rely on feedback from one
another and are not influence by top-down processing. Each stage represents independent
We’ve discussed lexical access in earlier lectures so let’s move to the second stage of sentence
comprehension, syntactic analysis. At this stage we use a strategy called parsing; when we parse
a sentence we are breaking it down into individual words and assigning each word a syntactic
We can either parse a sentence online, where we actively construct word trees as we’re
listening, or we can parse offline, which is when we wait for the sentence to be finished before
Ambiguity in sentences is a way of examining what methods we use. A garden path sentence for
example is a sentence that misleads us while we are processing online. These sentences typically
have an added noun phrase that trip us up.
o I.e. The girl gave the letter to her boyfriend to the postman.
Two other forms of ambiguity are lexical and syntactic ambiguity.
Lexical ambiguity for example would be “the girl saw the bug in the tree” where the word bug
can mean insect or some sort of recording device. Two ways to solve this would be to use
contextual information or frequency. The context would indicate