In this squib I point out a hitherto unnoticed interaction between weak crossover (WCO) and antecedent-contained deletion (ACD) that pro- vides another piece of evidence for an LF A ̄ -movement account of ACD and against the A-movement account proposed by Hornstein (1994), adding to the case against Hornstein’s proposal made by Ken- nedy (1997). Interestingly, the key piece of evidence, an ACD con- struction within a definite DP, implies that a quantificational type must be optionally available for definite DPs, but is exploited only when necessary to repair ACD violations.