Exploring the Guessing-Game Experimental Paradigm: Inferences From Closed- Versus Open-Ended Semantic Space

Area of Science:

• Cognitive Science
• Animal Communication
• Linguistics

Background:

• Previous studies suggest humans can interpret novel animal vocalizations and gestures from limited choices.
• The methodology used in prior research may influence observed comprehension levels.

Purpose of the Study:

• To investigate how different response formats and evaluation methods affect conclusions in signal comprehension studies.
• To determine the factors influencing human ability to understand novel vocalizations and ape gestures.

Main Methods:

• Replicated previous experiments using open-ended responses instead of multiple-choice.
• Applied exact matching, graded similarity ratings, and computational semantic similarity to evaluate vocalization data.
• Used graded similarity ratings for gesture data.

Main Results:

• Comprehension success varied significantly based on evaluation method: low for exact matching, moderate for similarity ratings, and high for computational measures.
• Signal properties (semantic category, transparency) were primary determinants of success, not individual participant differences.
• Participants could distinguish broad categories but struggled with specific concepts.

Conclusions:

• Open-ended response formats and diverse evaluation scales provide complementary insights into semantic understanding.
• Experimental design and evaluation methods are critical theoretical choices influencing the interpretation of signal comprehension.
• Findings challenge previous conclusions by highlighting the nuanced nature of interpreting novel signals.