In English, the NP in the genitive is the specifier of D' under which the genitive 's head subsumes.

E.g.: [DP [NP [N' [N Ali]]][D'[D 's][NP [N' [N mother]]]]]

enter image description here

In Malay, nouns in the genitive case aren't marked. Instead, one of the ways possession is denoted is through the position of nouns in the NP. Nouns that are placed consecutively may be in the genitive.

Malay: Ibu          bapa         Ali
Lit:   mother (GEN) father (GEN) Ali
Eng:   Ali's father's mother

The example above is not to be confused with "ibubapa Ali", which means "Ali's parents".

The main problem is the (GEN) terminal. How should nouns in the genitive without explicit genitive case marking be represented in X-bar tree diagrams? Is it plausible to have (GEN) to make clear the genitive feature of the noun?

This is what I propose, but I'm not sure if it would be accepted by or is theoretically sound for other linguists:

enter image description here

Please disregard the numbering of the nodes.

  • @IvanKapitonov, the head noun of a genitive of this kind should be the deepest in the NP structure, as is "mother" in the English translation. As parsed, "ibu" (mother) is the deepest in the NP, but in the left-most branch. This is because the NP is left-branching. This kind of Malay genitive has the possessor noun be at the right and the possessed noun at the left. Jun 9 '14 at 2:03
  • Oh, right. I'm removing my note. Jun 9 '14 at 8:36

No, that analysis can hardly be defended in terms of X-bar theory. The analysis shows D'2 lower in the structure than D2, and DP2 as a sister of the head D2. A projection of a head can never appear lower in the structure than the head, nor can it appear as a sister of the head, but rather it must appear above the head.

There are a couple of points that should be kept in mind when analyzing the structure of NPs. Perhaps the most controversial issue concerns whether noun phrases are NPs or DPs. The question assumes DPs, but that is a controversial assumption. A number of frameworks (Head-Driven Phrase Structure Grammar, Construction Grammar, Categorial Grammar, Meaning-Text Theory, etc.) assume NPs, not DPs. Hence one could reject the analysis shown based on this controversy alone. For arguments for and against DPs, see this article: https://en.wikipedia.org/wiki/Determiner_phrase.

My personal view is that the traditional NP analysis of noun phrases is more defensible, and I think that the example illustrates a difficulty with the DP analysis in general. Most discussions of the DP-hypothesis remain with single determiners; they do not explore stacked possessives like in the example in the question. Consider the following examples in this regard:

 1 a. Ali's father
   b. Ali's father's mother

If the 's is the head of the phrase Ali's father in (1a), then we should assume that the first 's is also the head of Ali's father's mother (not the second 's) in (1b). But intuitively that does does not seem right.

I think the traditional NP analysis is more consistent with such data, and in fact there is an empirical argument that supports the NP analysis, i.e proform substitution:

 a. Ali 's father
 b.  his   father

 a. Ali's father's mother
 b.  his  father's mother
 c.     his        mother

Proform substitution is a test for constituent structure that is widely employed in beginning textbooks for syntax. If a definite proform can be substituted in for a string of words, then that string is likely a constituent. In this case then, his can be substituted in for Ali's, indicating that Ali's is a constituent. Similarly, his can be substituted in for Ali's father's, indicating that Ali's father's is also a constituent.

Consider these facts with respect to the first DP tree in the question. That tree shows Ali's as a non-constituent, hence my conclusion is that the first tree in the question is also incorrect. The traditional NP-analysis, which would show Ali's as a constituent, is more defensible.

Concerning the second tree in the question, I think the constituent structure shown would be accurate for the NP Ali's father's mother, because the combination Ali's father's would receive the status of a constituent. But the labeling in that tree is all off.

  • I see what you mean and agree that D'2 being lower than D2 in the structure is totally wrong. The numbering isn't really of my doing but was what the parser automatically generated. Please disregard the numbering of the nodes, just not to digress from the main question in hand. I also agree that the NP analysis is much simpler and more widely accepted than the DP analysis, but just for the sake of varying my points of view, I used the DP analysis here. The main problem is the (GEN) terminal. Is it plausible to have them to make clear the genitive feature of the noun, even in the NP analysis? Jun 5 '14 at 5:17
  • 2
    @Adam, one thing to keep in mind is that the designation GEN denotes a morphological case. Malay, however, may lack case entirely, which would mean that from a technical point of view, talking about a GEN in Malay could be inaccurate. In any case, Malay is indeed indicating possession and it is apparently doing so by word order alone. I would simply position the possessor noun as a dependent of the possessed noun and be done with it. Struggling with an X-bar analysis is difficult, more difficult than its worth, I think. Jun 5 '14 at 5:30
  • @Adam, that said, it is of course good practice to think about such analyses, because it increases one's understanding of syntax theory in general. On an NP analysis, the possessor would indeed be terminal insofar as it would not be head over the noun. You'd have the following constituent structure for the more complex example: [[[[[Ali]'s] father]'s] mother]. Jun 5 '14 at 5:36
  • I understand that Indonesian uses prosody to mark this kind of thing, so maybe Malay does too? Jun 5 '14 at 12:10
  • 1
    We're talking about two contiguous nouns, which is the most common way of indicating possession, with or without morphological mortar. Why would anyone think counting angels on pinheads could explain this?
    – jlawler
    Jun 5 '14 at 14:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.