<div dir="ltr">I've got something I just found in Dorsey which may help back up des Herrn Professor Doktor Boyle claim that those pesky wa- prefixes are not valence reducers but actual arguments. Look at the agreement here:<br>
<br>(Dorsey 1890: 120.4-5)<br>Xubái égaⁿ égithaⁿi ki wébahaⁿ-hnáⁿi he.<br>sacred.3PROX 3.SIM <a href="http://say.to.PL">say.to.PL</a> when WA.know-FREQ.3PROX DECL.F<br>"Since he is sacred, when they say it to [one another], he always knows it of them."<br>
<br>It's important to realise that in O&P (other languages too?) "know" is a subject-object-raising verb, and obligatorily takes as its object the subject of the subordinate clause. (This is as far as I'm aware, I don't know if that's universally true of course.) The subject of the subordinate clause here is "indefinite-plural-human", just like the non-referring 3rd-person-plural stuff you get in Romance languages. And it just so happens that there is an object morpheme for that sort of argument: wa!<br>
<br>Oh well, the valence-reducer idea was nice though, wasn't it?<br><br>- Bryan<br>
</div>