Inductive Dependency Parsing

Philologists assure us that it’s worth learning ancient Greek just to read Homer. For any linguist, it’s definitely worth learning French, just to read Lucien Tesnière’s Elements de Syntaxe Structurale (Tesnière 1959). For any serious dependency parsing student or professional, it would have been worth learning Swedish, just to read Joakim Nivre’s Inductive Dependency Parsing if Nivre had not done the world the immense favor of writing his book in English, rather than in his native tongue of glory and heroes. The book demonstrates Nivre’s impressive ability to explain dependency grammar and dependency parsing clearly and succinctly to a wide audience. His overviews of dependency grammars (Section 3.1) and dependency parsing (Section 3.2) should be required reading in any course on the topics. He concisely sums up each relevant approach, often in a single sentence, and those wishing to learn more can rely on his extensive collection of references. The remainder of Chapter 3 sets the board for the proposed novel MALT parser—the book’s main innovation—by formalizing the dependency parsing task and specifying a particular type of deterministic shift-reduce parsing algorithm. Here, the set of parser actions are appropriate for dependency parsing, rather than for constituent parsing, and it is assumed that there is a one-to-one correspondence between dependency analyses and valid, complete sequences of parser actions. Nivre proves that the parsing algorithm correctly performs the formalized dependency parsing task, producing an acyclic, single-headed, projective dependency graph. Whereas the short Chapter 1 consists of the obligatory introduction, motivation, and book overview, it is interesting to discuss Nivre’s ambitious four criteria for a successful dependency parser, stated in Chapter 2: