A generalized four-step inference method for fuzzy quantified and truth-qualified natural language propositions

We propose a generalized inference method for constructing natural language communication. The method is used to obtain fuzzy quantifier Q\´ when “QA are F is τ ⇒ Q\´(m\´A) are mF is m\´\´ τ " inferred (Q, Q\´: fuzzy quantifiers, A: fuzzy subject, m, m\´, m": modifiers, F: fuzzy predicate, τ: truth qualifier). We show that Q\´ is resolved step by step for a non-increasing type (few,...) and a non-decreasing type (most,...).