The paper discusses the CoNLL-X shared task on Multilingual Dependency Parsing, which aimed to standardize and evaluate dependency parsing systems across multiple languages. The authors describe the process of converting treebanks for 13 languages into a unified dependency format and the evaluation metrics used. They provide an overview of the parsing approaches taken by participants and the results achieved. The paper also analyzes the factors that make certain languages or treebanks easier or harder to parse, such as the size of the training data, the proportion of new form and lemma values in the test set, and the ratio of coarse to fine-grained part-of-speech tags. The results show that Japanese is the easiest language to parse, while Turkish is the most difficult. The paper concludes with suggestions for future research, including the impact of annotations like lemmas and morphological features on parsing performance.The paper discusses the CoNLL-X shared task on Multilingual Dependency Parsing, which aimed to standardize and evaluate dependency parsing systems across multiple languages. The authors describe the process of converting treebanks for 13 languages into a unified dependency format and the evaluation metrics used. They provide an overview of the parsing approaches taken by participants and the results achieved. The paper also analyzes the factors that make certain languages or treebanks easier or harder to parse, such as the size of the training data, the proportion of new form and lemma values in the test set, and the ratio of coarse to fine-grained part-of-speech tags. The results show that Japanese is the easiest language to parse, while Turkish is the most difficult. The paper concludes with suggestions for future research, including the impact of annotations like lemmas and morphological features on parsing performance.