On Sat, Feb 18, 2017 at 11:11:48AM +0000, Anthony Appleyard wrote:
Computer-analyzing the 4 verb forms translating "wIlegh wIleghlu' nulegh nuleghlu'":- produces (1) wIlegh may mean: -- PP:we/[him|it] V:see :: we see him/it
(2) wIleghlu' may mean:- - PP:we/[him|it] V:see VS5:indef_subject
(3) nulegh may mean:- - PP:[he|it|they]/us V:see :: he/they see us
(4) nuleghlu' may mean:- - PP:[he|it|they]/us V:see VS5:indef_subject and it would be expected that (4) means "one sees us" and that (2) contradicts itself about what the subject is. But TKD says as a special rule that (2) means "one sees us", "we are seen", and says nothing directly about (4).
If not directly stated by Marc, it is widely assumed that *{nuleghlu'} would be agrammatical. When I made my comments about the Klingon passive being analogous to that of Finnish, I had forgotten about the fact that the "passive" verb agrees with its object using the subject side of the verbal prefixes (i.e. "I am seen" would be *vIleghlu'* rather than *muleghlu'* or something else). You could argue that, at some level, this means that Klingon sees the sole argument of a passive verb as a sort of subject. However, given that the prefix trick exists, I'm inclined to propose something different: perhaps the "subject" and "object" dimensions of the verb prefix don't actually correspond to the subject and object, but rather to the most prominent and second-most prominent arguments of the verb (I'm not sure if "prominent" is the best word to use here, but I'm going to stick with it for now, since I don't want to get sidetracked hunting down the correct terminology). The prominence hierarchy would look like this ("Null S" is the null subject of the passive construction with *-lu'*): 1st/2nd S > 3rd S > 1st/2nd DO > 1st/2nd IO > 3rd DO > 3rd IO > Null S The two highest non-oblique arguments are marked on the verb using the "subject" and "object" dimension of the agreement prefix. Thus, we get the regular conjugation of transitive verbs (the subject gets the "subject" dimension, and the direct object the "object" dimension), as well as the prefix trick (the subject gets the "subject" dimension, the first or second person indirect object the "object" dimension, and the third-person object remains unmarked on the verb) and now the passive construction (the direct object gets the "subject" dimension, and the null subject, which takes third-person agreement, gets the "object" dimension). Note that intransitive verbs, and verbs marked as if they were intransitive, aren't really part of this system, since they only deal with a single argument. And if you squint in just the right way, it feels sort of like the null marking in the third person (except for *lu-*) might be a logical consequence of the lower position of 3rd person arguments in this hierarchy. Do we know if the prefix trick can be used with a verb that takes *-lu'*, and if so, how exactly it works? If it can be, it might be necessary to adjust the above hierarchy slightly.
This reminds me of what happens in Celtic (where -r means "indefinite subject") compared to what happens in Latin (where "videmus" = "we see" and "videmur" means "we are seen".) In Latin, adding the "-r" makes the "-mus" change from specifying the one who sees to specifying the one who is seen, and that tense is called a passive, with "we" in subject position but meaning the one who is seen. This looks like that wIleghlu' is partly describable as a passive, making leghlu'wI' possible for "one who is seen".
Could you elaborate on this a bit? I'm not really following. - SapIr