Language & Cognition

Language & Cognition

Introduction

A level of language processing which is of great interest to psychologists is semantics - meaning. The mental lexicon is “the mental dictionary of words and their meanings”. On recognizing a word, access to its meaning is automatic - hence the stroop effect (people read the word red rather than naming the color if red is written in green ink) and the effect of word associations on word identification times (people are faster to say doctor is a word if it is preceded by nurse). In fact words are not the basic units of meaning.

Discussion

A morpheme is the “smallest unit of language that has meaning”. Thus cars has two morphemes - car and s (car is a free morpheme; s is bound). Word meaning is not quite the same as dictionary meaning. If one looks up a dictionary for a word, they will find a definition - one's lexical entry for a word will have similar information e.g. to sell means to exchange goods (usually for money). However, it is also know that it is usually people who sell things, and that selling happens at a certain place, time, etc. That is, one's lexical entry for the word will contain more information than one would find in a dictionary definition. Just how much semantic information is accessed when one recognizes a word is the question that arises here. Not all information - and it is context dependent. Piano can be seen as a very heavy object or as a musical instrument and the type of information accessed will in part depend on context. Garnham (2001) showed that the meaning accessed for cooked depends on what is cooked! One problem for the language processor is that many words are ambiguous - they are polysemous (having more than one meaning). For example, bank can be an institution, a piggy bank, or the side of a river. There has been much debate and a good deal of research on how we resolve the meaning of an ambiguous word. For example does an individual access all meanings immediately and then select an appropriate one, or does the context limit the availability of meanings from the offset. Green & Abutalebi (2008) used cross-modal priming and asked participants to listen to homonyms e.g. bugs embedded in different sentences. They were asked to make a lexical decision about a visual target which was related to only one of the meanings. Both related targets were primed i.e. ant and spy. When the experiment was repeated with a bigger delay between prime and target only the contextually relevant meaning was activated. This appears to support the autonomous view - that context only plays a role after the lexicon have been accessed. Later experiments are based on ERP research as well as this sort of priming study. (Goodglass et. al (2001) showed this measure is more sensitive to lexical ambiguity (as with 'bank' than standard priming studies and while their priming studies showed no early difference ...