Tema 7 – Structure and organization

Tema 7 – Structure and general organisation of the (mental) lexicon.

Linguistic models of the lexicon:

They assume that the lexicon is organized on the basis of lexical relations: synonymy, antonymy, class—inclusion (hyponymy), semantic fields, etc.

But note that we are interested in the mental lexicon (ML), so…

• Do those lexical relations have psychological validity?
• Could there be a different kind of organization in the ML?
• How can we know?

In the study of the ML, we assume:

1. Words in the ML must be organized in some way. Why?

• It is believed than an average speaker stores about 40.000 words in his/her ML.
• We retrieve and use lexical items at a rate of 2-3 per second.
• If words in the ML weren’t organized some way, this would be impossible.

So, how are words organized in the ML then?

• Hypothesis 1. They are listed alphabetically just like a desk dictionary. But this can’t be true! Because illiterate people use lexical items efficiently. And speakers of languages without writing systems use words as efficiently. So, we have to formulate a different hypothesis.
• Hypothesis 2. We can use the following evidence:
  • Speech errors or slips of the tongue:

They show normal lexical behavior. They are made by both educated and uneducated speakers. (so, you shouldn’t feel ashamed)

Two basic types:

• • • Semantic errors based on semantic relation, (word substitutions), you had one concept in mind but instead of articulating it you articulate a different

I wonder who invented crosswords > intrusive item (jigsaws > target item)

He came tomorrow (yesterday) < co-hyponyms

• • • Sound errors, not semantic relation, but phonological relation between both items, when several phonemes coincide between them.

The emperor had two porcupines (concubines)

There were lots of little orgasms (organism)

• • Speech disorders

Difficulties in speech production due to certain diseases.

Aphasia: speech impairment due to stroke, injury, etc. Some deal with problems in production and others in comprehension:

Examples of word-finding problems in aphasic patients:

• Chair for table (lexical relation)
• Knee for elbow (lexical relation)
• Hair for comb
• Tubber for butter (phonological relation)
• Leasing for ceiling (phonological relation)

Broca’s and Wernicke Area: Depending on the part some have problems with production and others with comprehension: Broca’s is involved with speech production and Wernicke’s is involved with linguistic comprehension (fluent speech but with no meaning).

Psycholinguistic experiments: ways in which we can gain information about the organization of the mental lexicon.

Experiments:

We put speakers to the test. A trivial example or word association: “Give me the 1st word you think of when I say X”.

The strongest links are usually:

• Co-hyponyms: black / white; salt / pepper
• Hyperonyms: red / color
• Antonyms: big / small; low / high
• And also, collocations (salt water) or simply frequent combinations.

So, this evidence together tells us: words are related in the ML phonologically (sound errors), semantically (through lexical relations like hyponymy, collocations, meronymy, etc.) and syntactically (in a corpus containing 200 speech errors Fay and Cutler (1977) showed that in 99% of the cases the intrusive word was of the same syntactic category as the target word).

Conclusion: the ML is a complex network of relations among lexical items. The ML has two modules: a phonological one and a semantic-syntactic one.

• • Psycholinguistic experiments

So, this evidence together tells us:

• Words are related in the ML phonologically semantically and syntactically.

Syntactically!

• In a corpus containing 200 speech errors Fay and Cutler showed that in 99% of the cases the intrusive word was of the same syntactic category as the target word.
• Conclusion:
  • The ML is a complex network of relations among lexical items.
  • The ML has two modules: a phonological one and a semantic-syntactic one.