Are the degree-distribution exponent and small-world structure of global syntactic dependency networks universal across UD languages?

For a language, build the global syntactic dependency network: nodes are word types (lemmas) and an undirected edge joins two words if, in some sentence of the corpus, one is the syntactic head of the other. Ferrer-i-Cancho, Sole & Kohler (2004) reported, for a few languages, that such networks are small-world and scale-free with degree distribution $P(k) \sim k^{-\gamma}$, $\gamma \approx 2.2$, disassortative, and hierarchical. QUESTION: over Universal Dependencies (100+ languages), estimate $\gamma$ with a principled power-law fit (and a model comparison against lognormal), together with the clustering coefficient, mean shortest-path length (small-world coefficient), and degree assortativity; report the cross-linguistic distribution of each statistic and test whether $\gamma \approx 2.2$ is universal or varies systematically with language type.

Acceptance. FULLY RESOLVES: for $\geq 50$ UD languages, build the global lemma-level syntactic dependency network and estimate $\gamma$ by Clauset-Shalizi-Newman MLE with a goodness-of-fit test and a likelihood-ratio comparison against a lognormal alternative, plus the clustering coefficient, mean path length (small-world coefficient $\sigma$), and degree assortativity; report the cross-linguistic distribution of each and assess whether $\gamma \approx 2.2$ is universal; ship a reproducible pipeline from public UD. PARTIAL: $\geq 5$ languages, or a rigorous power-law-vs-lognormal model comparison of one language's degree distribution.