Conal directs me towards a wonderful paper from J.T. Thomson in 1873, Philological Considerations on the Whence of the Maori, which is particularly relevant to the recent rates of word evolution work, and manages to beat lexicostatistics and Morris Swadesh by a good 75 years:
Primary words, i.e., those that express first wants in men in their infancy—and, equally so, tribes or nations in their infancy—are the most tenacious of existence. These are common nouns, pronouns, and verbs, but more particularly the first—such as man, woman, son, daughter, food, fruit, fish, etc.; or, I, you, he, we, etc.; or, go, come, give, kill, etc. In elucidating a subject such as this, therefore, we apply our enquiries to primary terms, which we may denominate as the fossils of the languages, so that we may, from their coincidence or approximations in different and distant communities, weigh the affinities of race or blood in the communities themselves.
But while primary words are the most lasting, yet they even are subject to slow and gradual change as ages roll on.
..and even more lexicostatistics-y:
Reverting, then, to the glossarial branch of the subject, in order to fairly weigh the respective affinities of the different races under review, as read by language, I must recall your attention to the fact stated in my former paper as to the relative number of primary words retained by an European language after eight hundred years of disconnection; these amount to only about one twenty-sixth of the whole. Mr. John Crawford, by his investigations, has declared that one fifty-seventh of the Malagasi and one-fiftieth of the Maori dictionaries were Malay, thus proving a connection whose intimacy on European experience can be approximately calculated. But I may venture to remark, from my own enquiries on the same subject, that had the above ethnographer or myself had the advantage of a critical knowledge of both or all languages, instead of only one (the Malay), double the equivalents might be found, and the approaches thus drawn nearer by half. Thus, Crawford states that out of 8,000 Malagasi words he detected only 140 Malayan; while I, out of Griffiths’ grammar, containing certainly not more than 500 words, detected eighty, in words that have had preservation throughout the whole region.