If a particular target type can mark agreement in gender then in many languages it must. … This may be called “enforced” gender agreement. … In a nutshell, if an agreement target can agree then typically it must agree, even if the agreement controller lacks the appropriate features.
Standardized
IF an agreement target can agree [in gender, number, or whatever], THEN typically it must agree, even if the agreement controller lacks the appropriate features.
1. Enforced gender agreement creates a problem if the controller is not specified for a gender (e.g., when it is an infinitive or a clause rather than a prototypical NP – see #2007) or if greater specificity is forced on the speaker than is possible/desirable. Different strategies: neutral/default agreement form, unique neutral agreement form, evasive form, special form.2. Claim repeated in slightly different form: “If a language has targets which distinguish gender, then typically they must distinguish gender.” (Corbett 1991: 218).3. What about the suspension of agreement in different constructions of the same target words? E.g. German: Adjectives can agree in gender, but they don’t when predicative, nor when attributive but postnominal (pur-er Whisky, Whisky pur). Is this “neutral/default” agreement, taking the shape of the bare stem, or simply “non-agreement”? I’d find it hard arguing against the latter – despite its being at odds with yet another universal, #26/63, claiming an agreement preference for postnominal vis-a-vis prenominal adjectives.
1. Enforced gender agreement creates a problem if the controller is not specified for a gender (e.g., when it is an infinitive or a clause rather than a prototypical NP – see #2007) or if greater specificity is forced on the speaker than is possible/desirable. Different strategies: neutral/default agreement form, unique neutral agreement form, evasive form, special form.2. Claim repeated in slightly different form: “If a language has targets which distinguish gender, then typically they must distinguish gender.” (Corbett 1991: 218).3. What about the suspension of agreement in different constructions of the same target words? E.g. German: Adjectives can agree in gender, but they don’t when predicative, nor when attributive but postnominal (pur-er Whisky, Whisky pur). Is this “neutral/default” agreement, taking the shape of the bare stem, or simply “non-agreement”? I’d find it hard arguing against the latter – despite its being at odds with yet another universal, #26/63, claiming an agreement preference for postnominal vis-a-vis prenominal adjectives.