Personal pronouns with first person n and second person m have been claimed to be frequent in the native languages of the Americas, widespread there, and rare elsewhere, and thus to indicate genetic unity of Amerind. A controlled cross-linguistic survey shows that these pronouns have an extensive yet restricted geographical range limited to the western Americas, and that they recur (though not frequently) elsewhere around the Pacific rim. This distribution removes the strongest (and perhaps the only) evidence for genetic relatedness of Amerind. In addition, on statistical grounds the n:m paradigm fails as a diagnostic of genetic relatedness, though equally clearly it cannot be due to universals or random chance. Certain other linguistic features and one mitochondrial DNA lineage have much the same geographical and statistical distribution. Though the language families in which these features appear cannot be shown to be genetically related, the families have clearly had some shared history (the type and degree not precisely specifiable) in the distant past. The n:m pronouns reflect a single, datable, noninitial and nonterminal phase in the settlement of the Americas and are probably the best linguistic marker of that phase.

We argue against the prevailing view that metathesis is somehow less natural phonetically than other processes and distinguished by a relatively greater phonological motivation. We survey cases of consonant-vowel metathesis—both synchronic processes and diachronic changes—with the goal of understanding how metathesis sound changes arise. We identify two types of CV metathesis, with distinct synchronic properties and distinct historical origins, and we argue that the two types do have natural, phonetic bases and fundamental commonalities.

In my annual report for 2005, published in Language 82.2 (June 2006), I noted that the end of 2005 meant that I was well past ‘the halfway mark of my editorship’ and that I was therefore ‘on the downhill slope ... of my service’ (p. 466). The final part of that slope began in spirit for me just after the January 2008 LSA annual meeting, some six months ago as I write this. But it has really begun in earnest only in the past month, as my successor has (finally) been selected. This happy turn of events puts me in a position to begin to reflect somewhat on my own term of office.

It is a truism that people speak 'loosely'—that is, that they often say things that we can recognize not to be true, but which come close enough to the truth for practical purposes. Certain expressions, such as those including exactly, all and perfectly, appear to serve as signals of the intended degree of approximation to the truth. This article presents a novel formalism for representing the notion of approximation to the truth, and analyzes the meanings of these expressions in terms of this formalism. Pragmatic looseness of this kind should be distinguished from authentic truth-conditional vagueness.

A clitic occurring mainly in same-subject chaining constructions in Ingush (Northeast Caucasian) exhibits the following characteristics in terms of Klavans's 1985 clitic typology: it is positioned with respect to the final element of its domain, before that element, and is enclitic on a preceding element. The clitic therefore provides a good example of Klavans's Type 5 clitic, the existence of which is disputed. The search for evidence bearing on the status of the clitic also results in a more comprehensive treatment of clause chaining in Ingush than has previously been available.