Postposition ה- notation in construct chains

[Susan]

I'm trying to understand the "directional" postposition ה-. I understand that the noun to which it is attached is generally its complement, and the ה itself is usually either an adjunct or a complement depending on the semantics of the verb on which it elaborates. Most of them are straightforward, but in the case where the noun is in construct, there seem to be two different ways of diagraming the construction and I'm trying to figure out the difference between them.

 

The first is what I would expect, with the nomen regens (= bound noun) subordinated to the directional suffix as a complement; the nomen rectum (= clinic host) is an adjunct to the bound form. For instance, Gen 42:29:

 

 

The second seems to be more common, where the ה is a "connector". For instance (using the same phrase), Gen 45:17:

 

 

Here I gather the superscripted + and the & indicate a compound complement. The host is still subordinated as an adjunct in the way I would expect, but I don't think the bound form is subordinated to the ה, and I can't figure out why the latter is considered a "connector".

 

I wondered if the deciding factor was whether the entire phrase was a complement (as in Gen 45:17) (not that I understand why that would matter), since this is most common role for ה phrases (it seems to me), but Ex. 10:19 is a complement and looks more like the Gen 42:29 example:

 

 

(though here the host is actually a level up compared to the prior example, i.e. subordinated to the ה and at the same level as the bound form; at least the bound form and its ה are doing what I would expect).

 

I'm wondering if someone can help me figure out:

 

1. what determines whether the phrase is considered compound, with the ה as a "connector" (i.e. the difference between #1 and #2) , and 

2. what determines whether the nomen rectum is subordinated (directly) to the ה or to the nomen regens (i.e. the difference between #1 and #3).

 

Thanks!

[Robert Holmstedt]

Susan,

 

The Exod 10.19 (and 5 others in the database) were simple errors in the tagging formula.

 

Gen 42.29 and 45.17 are tagged *exactly* the same way. The difference in the trees is that Accordance is mistakenly reading the structure in 45.17 as a compound (whereas in 42.29, since that whole phrase is an appositive, it hasn't tricked Accordance like the 45.17 structure). I'll report this is a minor error.

 

1. The ה is never a "connector" creating a compound. 

2. Because the postpositive ה obscures the bound morphology of the word it attaches to, determining whether this word is bound to the following NP is based on whether the second NP has any other more likely role in the larger clause. If not, it is likely the clitic host for the first NP. 

[Susan]

Thanks, everything makes sense now.