Semantic role labeling (SRL) identifies predicate-argument structure(s) in agiven sentence. Although different languages have different argumentannotations, polyglot training, the idea of training one model on multiplelanguages, has previously been shown to outperform monolingual baselines,especially for low resource languages. In fact, even a simple combination ofdata has been shown to be effective with polyglot training by representing thedistant vocabularies in a shared representation space. Meanwhile, despite thedissimilarity in argument annotations between languages, certain argumentlabels do share common semantic meaning across languages (e.g. adjuncts havemore or less similar semantic meaning across languages). To leverage suchsimilarity in annotation space across languages, we propose a method calledCross-Lingual Argument Regularizer (CLAR). CLAR identifies such linguisticannotation similarity across languages and exploits this information to map thetarget language arguments using a transformation of the space on which sourcelanguage arguments lie. By doing so, our experimental results show that CLARconsistently improves SRL performance on multiple languages over monolingualand polyglot baselines for low resource languages.