When we consider the question categories, various taxonomies were proposed. An important…
- User
When we consider the question categories, various taxonomies were proposed. An important contribution towards this direction is Lehnert’s classification [18]. As part of development of a computational model for question-answering, Lehnert classified questions based on the idea of conceptual categorization. As per this idea, in order for the question to be interpreted correctly, it must be placed in the right conceptual category, otherwise it will lead to wrong reasoning. In this sense, the emphasis should be on the context in which the question was asked. Accordingly, Lehnert proposed thirteen such conceptual categories, namely causal antecedent, goal orientation, enablement, causal consequent, verification, disjunctive, instrumental, concept completion, expectational, judgmental, quantification, feature specification and request
- Assistant
You're correct. Wendy Lehnert's classification system for questions has been highly influential in the field of AI and natural language processing. Here is a brief rundown of what each category means:
- Causal Antecedent: Questions that ask about the cause of an event or situation.
- Goal Orientation: Questions that ask about the purpose or goal of an action or decision.
- Enablement: Questions that ask about the necessary conditions to make an event happen.
- Causal Consequent: Questions asking about the results or effects of an event or situation.
- Verification: Questions that seek to confirm the truth or false of a statement.
- Disjunctive: Questions that present alternatives and ask the user to choose one.
- Instrumental: Questions about the tools or methods used to accomplish something.
- Concept Completion: Questions that require filling in a missing piece of information or details.
- Expectational: Questions about what is likely to happen in a given scenario.
- Judgmental: Questions that ask for an opinion or evaluation.
- Quantification: Questions that ask about the amount, number, or degree of something.
- Feature Specification: Questions that ask for specific attributes or characteristics of something.
- Request: Questions that ask the addressee to perform an action.
These categories can be helpful in understanding the intent behind a question and guiding the system to provide more accurate and relevant responses.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.