Measuring number of roots as an index of a language's maturity... what could possibly go wrong?

Lojban: >16500 (much of which is proper names)
Esperanto: >15000 (possibly much larger)
Verdurian: 6450
Quenya: ~4600 published (Ardalambion)
Sindarin: ~2700 published (Hiswelókë)
Láadan: ~1800 (incl. numerous affixes and other artefacts)
Ithkuil: 900 published, 3600 claimed