Are lexical representations graded or discrete?

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Lexical Semantics

Background:

• Traditional models propose discrete mental lexical representations (lemmas) corresponding to distinct word meanings.
• Homophones (e.g., 'bat') are thought to have separate lemmas for unrelated meanings, while polysemes (e.g., 'paper') share a single lemma for related senses.
• However, cognitive processes are often graded, prompting investigation into whether lemmas are also graded.

Purpose of the Study:

• To explore whether mental lexical representations (lemmas) are discrete or graded.
• To examine the semantic processing of homophones and polysemes using a picture-word interference paradigm.
• To test the multiple-lemma versus core-lemma account of polysemy.

Main Methods:

• A preregistered picture-word interference study was conducted.
• Participants named pictures, with semantic competitors presented for non-depicted meanings of homophones and polysemes.
• Naming times were analyzed to infer lemma structure based on facilitation or inhibition effects.

Main Results:

• Semantic competitors to non-depicted senses of both homophones and polysemes unexpectedly facilitated picture naming.
• This facilitation effect, observed across varying degrees of semantic relatedness, contradicts predictions of discrete lemma models for polysemes.
• Results did not definitively distinguish between graded and discrete lemma structures but provided evidence for polysemy.

Conclusions:

• The findings support a multiple-lemma account for polysemes over a core-lemma account.
• The study challenges the assumption of a single, shared lemma for related senses of polysemous words.
• Further research is needed to fully elucidate the graded versus discrete nature of mental lexical representations.