Searching trademark databases for verbal similarities

We review various issues related to techniques for searching through collections of trademarks for phonetic and other verbal similarities. Based on recent case studies, we survey judicial arbitrations that have established whether trademarks are similar or not and comment on popular algorithms for performing approximate searches, both phonetically and based on letter sequences. The particularities of international registrations and multilingual trademarks are also discussed. Overall, fuzzy searching for verbal similarities is a difficult task that calls for a cautious attitude. No solution can guarantee a fully automatic procedure. A manual validation of results will always need to be performed.